We now cache the negative outcome of PreferredConstructor.isConstructorParameter(…) to avoid iterations and iterator allocations which roughly improves typical converter usage by roughly 32%.
Before:
Benchmark Mode Cnt Score Error Units
TypicalEntityReaderBenchmark.simpleEntityGeneratedConstructorAndField thrpt 10 6848447,474 ± 554354,377 ops/s
After
Benchmark Mode Cnt Score Error Units
TypicalEntityReaderBenchmark.simpleEntityGeneratedConstructorAndField thrpt 10 9071099,879 ± 1423166,087 ops/s
Closes#2295.
We now leniently skip parameter name resolution for types using unsigned types. The issue is caused by Kotlin's DefaultConstructorMarker that doesn't report a parameter name.
Closes#2215
Allow obtaining an EvaluationContext with a reduced set of extensions that are required for ExpressionDependencies instead of initializing all extensions.
We now use a dedicated GetOptions instance for the lookup of the parent path in case SetOptions is defined to skip nulls so that the lookup does not fail if the parent path contains a null value.
Use a more lenient approach, that allows metadata creation, when looking up persistent entities. This allows eg. a configured AuditingHandler to kick in without having to register an initial entity set in first place.
Original pull request: #462.
Use KotlinDetector instead of ReflectionUtils.isKotlinClass(…). Use replacement MethodParameter(…).withContainingClass(…) instead of GenericTypeResolver.resolveParameterType(…). Use MergedAnnotations.from(…).isPresent(…) instead of AnnotationUtils.isAnnotationMetaPresent(…).
kotlin.Unit is now considered a simple type to avoid PersistentEntity creation.
ReflectionUtils.isVoid(…) now encapsulates the check so that calling code doesn't need to check if Kotlin is on the class path.
We now reuse the new bean in InstantiationAwarePropertyAccessor when setting properties. Previously, we used the initial bean state as the bean was held by a delegate PersistentPropertyAccessor which caused only the last set property to be visible.
Always fall back to reflection when the system property for org.graalvm.nativeimage.imagecode is set.
This is required because on the fly code generation is not supported in this case.
Original pull request: #456.
Update classes so that each member is declared on its own line. This
help to make them slightly easier to see when not using an IDE.
Original pull request: #448.
Replace the few occurrences where SLF4J was being used directly so that
all logging now happens via commons-logging. The log patterns used in
SLF4J messages have been replaced with `LogMessage` which was introduced
in Spring Framework 5.2.
Original pull request: #448.
Replace the use of an anonymous class with a static initializer block.
This update also changes the constant to an unmodifiable variant.
Original pull request: #448.
Rename static final `CACHE` members to use the lowercase form. Although
the caches are static final, they are not constant values.
Original pull request: #448.
Use mutated object as guard for synchronization. Copy discovered callbacks to cached callbacks.
Reduce concurrency in unit test to reduce test load. Guard synchronization with timeouts.
Original pull request: #446.
Switch property parameter cache from HashMap and external locks to ConcurrentHashMap to improve multi-threaded locking behavior during isConstructorParameter(…) initialization.
Original pull request: #437.
Reorder constructor filtering with the goal to delay/avoid potentially expensive PersistenceConstructor creation with parameter lookup as much as possible.
The lookup is (in a default setup) only done once per domain type, but depending on the number of entities to inspect the change can help save some cycles for larger domains.
Before:
Benchmark Mode Cnt Score Error Units
EntityMetadataBenchmark.alwaysNew thrpt 10 224318,163 ± 42542,453 ops/s
After:
Benchmark Mode Cnt Score Error Units
EtityMetadataBenchmark.alwaysNew thrpt 10 420053,505 ± 9288,093 ops/s
public class DomainType {
private String id;
private @Id String theId;
private String firstname, lastname;
private Integer age;
public DomainType(String id, String theId, String firstname, String lastname, Integer age) {
this.id = id;
this.theId = theId;
this.firstname = firstname;
this.lastname = lastname;
this.age = age;
}
@PersistenceConstructor
public DomainType(String firstname, String lastname) {
this.firstname = firstname;
this.lastname = lastname;
}
}
Original pull request: #440.
We now detect whether callbacks are present before we create result lists instead of always allocating a Collection. CallbackRetriever now also returns a cached variant of its list of callbacks instead of always allocating a new ArrayList.
We now expose a dedicated PersistentPropertyPathAccessor and optionally take options for both the read and write access to properties. On read, clients can define to eagerly receive null in case one of the path segments is null. By default, we reject those as it indicates that there might be an issue with the backing object if the client assumes it can look up the more deeply nested path to receive a result.
On write access clients can define to either reject, skip or log intermediate null segments where skip means, that the setting is just silently aborted. Clients can also control how to deal with collection and map intermediates. By default we now propagate the attempt to set a value to all collection or map items respectively. This could be potentially extended to provide a filter receiving both the property and property value to selectively propagate those calls in the future.
The separation of these APIs (PersistentPropertyAccessor and PersistentPropertyPathAccessor) is introduced as the latter can be implemented on top of the former and the former potentially being dynamically generated. The logic of the latter can then be clearly separated from the actual individual property lookup. ConvertingPropertyAccessor is now a PersistentPropertyPathAccessor, too, and overrides SimplePersistentPropertyPathAccessor.getTypedProperty(…) to plug in conversion logic. This allows custom collection types being used if the ConversionService backing the accessor can convert from and to Collection or Map respectively.
Deprecated all PersistentPropertyPath-related methods on PersistentPropertyAccessor for removal in 2.3.
MappingAuditableBeanWrapperFactory now uses these new settings to skip both null values as well as collection and map values when setting auditing metadata.
Related tickets: DATACMNS-1438, DATACMNS-1461, DATACMNS-1609.
Added ClassUtils.ifPresent(…) to conditionally call back a Consumer<Class> if a class is available from the given ClassLoader. Extract isSuspend(…) method into KotlinReflectionUtils. Deprecate Kotlin-related methods in our ReflectionUtils as parts are available from Spring Framework directly.
Rename CoCrudRepository to CoroutineCrudRepository and CoroutineSortingRepository. Add tests for KotlinReflectionUtils to test calls without Kotlin dependencies.
Original pull request: #415.
We now ship and use a PropertyAccessor implementation by default that will fall back to use the persistence constructor and creating a new instance to "set" a property. This allows completely immutable types, ommitting previously needed wither methods, e.g. to populate automatically generated identifier values.
This is implemented by using the EntityInstantiator available and a ParameterValueProvider that returns the value of the properties of the current instance but replacing the one for the value to be set. This requires moving EntityInstantiators, the EntityInstantiator SPI as well as all existing implementations (reflective, class generating and the Kotlin one) into the mapping.model package. Legacy instances stay around in deprecated form to avoid breaking existing clients (other store modules as well as user implementations of EntityInstantiator).
The newly introduced implementations stay package protected as their sole users are now colocated in the same package. Access from the legacy types is handled via InternalEntityInstantiatorFactory, which is introduced in deprecated from for removal alongside the actual deprecations.
We now skip PersistentPropertyPath instances pointing to auditing properties for which the path contains a collection or map path segment as the PersistentPropertyAccessor currently cannot handle those. A more extensive fix for that will be put in place for Moore but requires more extensive API changes which we don't want to ship in a Lovelace maintenance release.
Related tickets: DATACMNS-1461.
PropertyAccessorClassGenerator now tries to look up the class to be generated first to potentially reuse an already existing one and avoid the recreation and registration of a same class which would trigger a Linkage error as a classloader cannot hold two classes with the same name.
The root of the problem is in the fact that the accessor instances are held in per instance caches in EntityInstantiators, so that multiple of those might try to create the same accessor class for a given domain type.
We now remove partially populated PersistentEntity instances from the cache held in AbstractMappingContext as the creation could fail for other RuntimeExceptions other than a MappingException and we wouldn't want to keep the instances around in any case.
Slight refactorings in the unit tests so that we can easily create MappingContext instances that reject certain types with certain exceptions.
PersistentPropertyAccessor now has an overloaded method to pass in a TraversalContext that can be prepared to property handlers for PersistentProperties. During the traversal, the handler is called with the value for the property so that it can e.g. extract list elements, map values etc.
Add dedicated interfaces for sync and reactive usage.
Hide default implementation and use reflective callback method lookup.
Update documentation and add initial reference documentation snippet for store specific modules.
Original Pull Request: #332
This is a draft for a possible Entity Callback API heavily inspired by Spring Framework's Application Event listeners. Entity callbacks are callbacks for entities that allow for entity modification at certain check points such as before saving an entity or after loading it.
Entity callbacks consist of a marker-interface for all entity callback types and the EntityCallbacks API to dispatch callbacks.
Entity callbacks are picked up from the ApplicationContext and invoked sequentially according to their ordering.
Usage example:
AnnotationConfigApplicationContext ctx = new AnnotationConfigApplicationContext(MyConfig.class);
SimpleEntityCallbacks callbacks = new SimpleEntityCallbacks(ctx);
PersonDocument personDocument = new PersonDocument(null, "Walter", null);
PersonDocument afterCallback = callbacks.callback(personDocument, BeforeSaveCallback.class,
BeforeSaveCallback::onBeforeSave);
Mono<PersonDocument> afterCallback = callbacks.callbackLater(personDocument, ReactiveBeforeSaveCallback.class,
ReactiveBeforeSaveCallback::onBeforeSave);
@Configuration
static class MyConfig {
@Bean
BeforeSaveCallback<Person> personCallback() {
return object -> {
object.setSsn(object.getFirstName().length());
return object;
};
}
@Bean
MyReactiveBeforeSaveCallback callback() {
return new MyReactiveBeforeSaveCallback();
}
}
class MyReactiveBeforeSaveCallback implements ReactiveBeforeSaveCallback<Person> {
@Override
public Mono<Person> onBeforeSave(Person object) {
PersonDocument result = new PersonDocument(object.getFirstName().length(), object.getFirstName(),
object.getLastName());
return Mono.just(result);
}
}
Original Pull Request: #332
