*/
package org.opendaylight.mdsal.binding.api;
+import com.google.common.util.concurrent.FluentFuture;
import org.eclipse.jdt.annotation.NonNull;
import org.opendaylight.mdsal.common.api.LogicalDatastoreType;
+import org.opendaylight.mdsal.common.api.TransactionDatastoreMismatchException;
import org.opendaylight.yangtools.yang.binding.DataObject;
import org.opendaylight.yangtools.yang.binding.InstanceIdentifier;
* whole subtree will be replaced by the specified data.
*
* <p>
- * This method does not automatically create missing parent nodes. It is equivalent to invoking
- * {@link #put(LogicalDatastoreType, InstanceIdentifier, DataObject, boolean)}
- * with <code>createMissingParents</code> set to false.
- *
- * <p>
- * If you need to make sure that a parent object exists but you do not want modify
- * its pre-existing state by using put, consider using {@link #merge} instead.
+ * If you need to make sure that a parent object exists but you do not want modify its pre-existing state by using
+ * put, consider using {@link #merge} instead.
*
* @param store the logical data store which should be modified
* @param path the data object path
* @param data the data object to be written to the specified path
* @throws IllegalStateException if the transaction has already been submitted
- * @throws NullPointerException if any of the arguments is null
+ * @throws NullPointerException if any of the arguments is {@code null}
+ * @throws TransactionDatastoreMismatchException if this transaction is already bound to a different data store
*/
<T extends DataObject> void put(@NonNull LogicalDatastoreType store, @NonNull InstanceIdentifier<T> path,
@NonNull T data);
/**
* Stores a piece of data at the specified path. This acts as an add / replace operation, which is to say that whole
- * subtree will be replaced by the specified data.
+ * subtree will be replaced by the specified data. Unlike
+ * {@link #put(LogicalDatastoreType, InstanceIdentifier, DataObject)}, this method will attempt to create
+ * semantically-significant parent nodes, like list entries and presence containers, as indicated by {@code path}.
*
* <p>
* If you need to make sure that a parent object exists but you do not want modify its pre-existing state by using
* put, consider using {@link #merge} instead.
*
* <p>
- * Note: Using <code>createMissingParents</code> with value true, may introduce garbage in data store, or recreate
- * nodes, which were deleted by previous transaction.
+ * <b>WARNING:</b> Using this method may introduce garbage in data store, or recreate nodes, which were deleted by
+ * a previous transaction. It also has a significantly higher cost than
+ * {@link #put(LogicalDatastoreType, InstanceIdentifier, DataObject)} and should only be used when
+ * absolutely necessary.
*
* @param store the logical data store which should be modified
* @param path the data object path
* @param data the data object to be written to the specified path
- * @param createMissingParents if {@link #CREATE_MISSING_PARENTS}, any missing parent nodes will be automatically
- * created using a merge operation.
* @throws IllegalStateException if the transaction has already been submitted
- * @throws NullPointerException if any of the arguments is null
+ * @throws NullPointerException if any of the arguments is {@code null}
+ * @throws TransactionDatastoreMismatchException if this transaction is already bound to a different data store
+ * @deprecated Use of this method is a manifestation of bad lifecycle management: it attempts to create data tree
+ * parent nodes which may have semantic meaning without assigning responsibility. The datastore handles
+ * all the cases which do not attach semantics, such as {@code container}s without {@code presence},
+ * {@code augmentation} and {@code list} encapsulation.
+ * This method does not work in the general case, where there are:
+ * <ul>
+ * <li>{@code container} parents with {@code presence}, as these form a {@code mandatory} enforcement
+ * boundary. We cannot infer the mandatory nodes from {@code path} and hence we may end up wanting
+ * to create an invalid structure</li>
+ * <li>{@code list} parents with {@code unique} encompassing other leaves than {@code key}. While we
+ * can re-create the key {@code leaf} items, we have no way of constructing of {@code unique}
+ * requirements.</li>
+ * </ul>
+ * Based on this analysis, all users of this method need to be migrated to have a proper lifecycle
+ * relationship with entities responsible for managing such semantic items which are created by this
+ * method.
*/
- <T extends DataObject> void put(@NonNull LogicalDatastoreType store, @NonNull InstanceIdentifier<T> path,
- @NonNull T data, boolean createMissingParents);
+ @Deprecated(since = "11.0.3")
+ <T extends DataObject> void mergeParentStructurePut(@NonNull LogicalDatastoreType store,
+ @NonNull InstanceIdentifier<T> path, @NonNull T data);
/**
* Merges a piece of data with the existing data at a specified path. Any pre-existing data which is not explicitly
* overwritten will be preserved. This means that if you store a container, its child lists will be merged.
*
* <p>
- * This method does not automatically create missing parent nodes. It is equivalent to invoking
- * {@link #merge(LogicalDatastoreType, InstanceIdentifier, DataObject, boolean)}
- * with <code>createMissingParents</code> set to false.
- *
- * <p>
* If you require an explicit replace operation, use {@link #put} instead.
+ *
* @param store the logical data store which should be modified
* @param path the data object path
* @param data the data object to be merged to the specified path
* @throws IllegalStateException if the transaction has already been submitted
- * @throws NullPointerException if any of the arguments is null
+ * @throws NullPointerException if any of the arguments is {@code null}
+ * @throws TransactionDatastoreMismatchException if this transaction is already bound to a different data store
*/
<T extends DataObject> void merge(@NonNull LogicalDatastoreType store, @NonNull InstanceIdentifier<T> path,
@NonNull T data);
/**
* Merges a piece of data with the existing data at a specified path. Any pre-existing data which is not explicitly
- * overwritten will be preserved. This means that if you store a container, its child lists will be merged.
+ * overwritten will be preserved. This means that if you store a container, its child lists will be merged. Unlike
+ * {@link #merge(LogicalDatastoreType, InstanceIdentifier, DataObject)}, this method will attempt to create
+ * semantically-significant parent nodes, like list entries and presence containers, as indicated by {@code path}.
*
* <p>
* If you require an explicit replace operation, use {@link #put} instead.
*
+ * <p>
+ * <b>WARNING:</b> Using this method may introduce garbage in data store, or recreate nodes, which were deleted by
+ * a previous transaction. It is not necessary in most scenarios and has a significantly higher cost
+ * than {@link #merge(LogicalDatastoreType, InstanceIdentifier, DataObject)}. It should only be used
+ * when absolutely necessary.
+ *
* @param store the logical data store which should be modified
* @param path the data object path
* @param data the data object to be merged to the specified path
- * @param createMissingParents if {@link #CREATE_MISSING_PARENTS}, any missing parent nodes will be automatically
- * created using a merge operation.
* @throws IllegalStateException if the transaction has already been submitted
- * @throws NullPointerException if any of the arguments is null
+ * @throws NullPointerException if any of the arguments is {@code null}
+ * @throws TransactionDatastoreMismatchException if this transaction is already bound to a different data store
+ * @deprecated Use of this method is a manifestation of bad lifecycle management: it attempts to create data tree
+ * parent nodes which may have semantic meaning without assigning responsibility. The datastore handles
+ * all the cases which do not attach semantics, such as {@code container}s without {@code presence},
+ * {@code augmentation} and {@code list} encapsulation.
+ * This method does not work in the general case, where there are:
+ * <ul>
+ * <li>{@code container} parents with {@code presence}, as these form a {@code mandatory} enforcement
+ * boundary. We cannot infer the mandatory nodes from {@code path} and hence we may end up wanting
+ * to create an invalid structure</li>
+ * <li>{@code list} parents with {@code unique} encompassing other leaves than {@code key}. While we
+ * can re-create the key {@code leaf} items, we have no way of constructing of {@code unique}
+ * requirements.</li>
+ * </ul>
+ * Based on this analysis, all users of this method need to be migrated to have a proper lifecycle
+ * relationship with entities responsible for managing such semantic items which are created by this
+ * method.
*/
- <T extends DataObject> void merge(@NonNull LogicalDatastoreType store, @NonNull InstanceIdentifier<T> path,
- @NonNull T data, boolean createMissingParents);
+ @Deprecated(since = "11.0.3")
+ <T extends DataObject> void mergeParentStructureMerge(@NonNull LogicalDatastoreType store,
+ @NonNull InstanceIdentifier<T> path, @NonNull T data);
/**
* Removes a piece of data from specified path. This operation does not fail if the specified path does not exist.
* @param store Logical data store which should be modified
* @param path Data object path
* @throws IllegalStateException if the transaction was committed or canceled.
+ * @throws NullPointerException if any of the arguments is {@code null}
+ * @throws TransactionDatastoreMismatchException if this transaction is already bound to a different data store
*/
void delete(@NonNull LogicalDatastoreType store, @NonNull InstanceIdentifier<?> path);
/**
- * Flag value indicating that missing parents should be created.
- */
- boolean CREATE_MISSING_PARENTS = true;
-
- /**
- * Flag value indicating that missing parents should cause an error.
+ * Return a {@link FluentFuture} which completes.
+ *
+ * @return A future which completes when the requested operations complete.
*/
- boolean FAIL_ON_MISSING_PARENTS = false;
+ @NonNull FluentFuture<?> completionFuture();
}
Code Review / mdsal.git / blobdiff