Develop Reference

Change the value of a parameter annotation [duplicate]

Imagine there is a class:
#Something(someProperty = "some value")
public class Foobar {
//...
}
Which is already compiled (I cannot control the source), and is part of the classpath when the jvm starts up. I would like to be able to change "some value" to something else at runtime, such that any reflection thereafter would have my new value instead of the default "some value".
Is this possible? If so, how?

Warning: Not tested on OSX - see comment from #Marcel
Tested on OSX. Works fine.
Since I also had the need to change annotation values at runtime, I revisited this question.
Here is a modified version of #assylias approach (many thanks for the inspiration).
/**
* Changes the annotation value for the given key of the given annotation to newValue and returns
* the previous value.
*/
#SuppressWarnings("unchecked")
public static Object changeAnnotationValue(Annotation annotation, String key, Object newValue){
Object handler = Proxy.getInvocationHandler(annotation);
Field f;
try {
f = handler.getClass().getDeclaredField("memberValues");
} catch (NoSuchFieldException | SecurityException e) {
throw new IllegalStateException(e);
}
f.setAccessible(true);
Map<String, Object> memberValues;
try {
memberValues = (Map<String, Object>) f.get(handler);
} catch (IllegalArgumentException | IllegalAccessException e) {
throw new IllegalStateException(e);
}
Object oldValue = memberValues.get(key);
if (oldValue == null || oldValue.getClass() != newValue.getClass()) {
throw new IllegalArgumentException();
}
memberValues.put(key,newValue);
return oldValue;
}
Usage example:
#Retention(RetentionPolicy.RUNTIME)
#Target(ElementType.TYPE)
public #interface ClassAnnotation {
String value() default "";
}
#Retention(RetentionPolicy.RUNTIME)
#Target(ElementType.FIELD)
public #interface FieldAnnotation {
String value() default "";
}
#Retention(RetentionPolicy.RUNTIME)
#Target(ElementType.METHOD)
public #interface MethodAnnotation {
String value() default "";
}
#ClassAnnotation("class test")
public static class TestClass{
#FieldAnnotation("field test")
public Object field;
#MethodAnnotation("method test")
public void method(){
}
}
public static void main(String[] args) throws Exception {
final ClassAnnotation classAnnotation = TestClass.class.getAnnotation(ClassAnnotation.class);
System.out.println("old ClassAnnotation = " + classAnnotation.value());
changeAnnotationValue(classAnnotation, "value", "another class annotation value");
System.out.println("modified ClassAnnotation = " + classAnnotation.value());
Field field = TestClass.class.getField("field");
final FieldAnnotation fieldAnnotation = field.getAnnotation(FieldAnnotation.class);
System.out.println("old FieldAnnotation = " + fieldAnnotation.value());
changeAnnotationValue(fieldAnnotation, "value", "another field annotation value");
System.out.println("modified FieldAnnotation = " + fieldAnnotation.value());
Method method = TestClass.class.getMethod("method");
final MethodAnnotation methodAnnotation = method.getAnnotation(MethodAnnotation.class);
System.out.println("old MethodAnnotation = " + methodAnnotation.value());
changeAnnotationValue(methodAnnotation, "value", "another method annotation value");
System.out.println("modified MethodAnnotation = " + methodAnnotation.value());
}
The advantage of this approach is, that one does not need to create a new annotation instance. Therefore one doesn't need to know the concrete annotation class in advance. Also the side effects should be minimal since the original annotation instance stays untouched.
Tested with Java 8.

This code does more or less what you ask for - it is a simple proof of concept:
a proper implementation needs to also deal with the declaredAnnotations
if the implementation of annotations in Class.java changes, the code will break (i.e. it can break at any time in the future)
I have no idea if there are side effects...
Output:
oldAnnotation = some value
modifiedAnnotation = another value
public static void main(String[] args) throws Exception {
final Something oldAnnotation = (Something) Foobar.class.getAnnotations()[0];
System.out.println("oldAnnotation = " + oldAnnotation.someProperty());
Annotation newAnnotation = new Something() {
#Override
public String someProperty() {
return "another value";
}
#Override
public Class<? extends Annotation> annotationType() {
return oldAnnotation.annotationType();
}
};
Field field = Class.class.getDeclaredField("annotations");
field.setAccessible(true);
Map<Class<? extends Annotation>, Annotation> annotations = (Map<Class<? extends Annotation>, Annotation>) field.get(Foobar.class);
annotations.put(Something.class, newAnnotation);
Something modifiedAnnotation = (Something) Foobar.class.getAnnotations()[0];
System.out.println("modifiedAnnotation = " + modifiedAnnotation.someProperty());
}
#Something(someProperty = "some value")
public static class Foobar {
}
#Retention(RetentionPolicy.RUNTIME)
#interface Something {
String someProperty();
}

This one works on my machine with Java 8. It changes the value of ignoreUnknown in the annotation #JsonIgnoreProperties(ignoreUnknown = true) from true to false.
final List<Annotation> matchedAnnotation = Arrays.stream(SomeClass.class.getAnnotations()).filter(annotation -> annotation.annotationType().equals(JsonIgnoreProperties.class)).collect(Collectors.toList());
final Annotation modifiedAnnotation = new JsonIgnoreProperties() {
#Override public Class<? extends Annotation> annotationType() {
return matchedAnnotation.get(0).annotationType();
} #Override public String[] value() {
return new String[0];
} #Override public boolean ignoreUnknown() {
return false;
} #Override public boolean allowGetters() {
return false;
} #Override public boolean allowSetters() {
return false;
}
};
final Method method = Class.class.getDeclaredMethod("getDeclaredAnnotationMap", null);
method.setAccessible(true);
final Map<Class<? extends Annotation>, Annotation> annotations = (Map<Class<? extends Annotation>, Annotation>) method.invoke(SomeClass.class, null);
annotations.put(JsonIgnoreProperties.class, modifiedAnnotation);

SPRING can do this job very easily , might be useful for spring developer .
follow these steps :-
First Solution :-
1)create a Bean returning a value for someProperty . Here I injected the somePropertyValue with #Value annotation from DB or property file :-
#Value("${config.somePropertyValue}")
private String somePropertyValue;
#Bean
public String somePropertyValue(){
return somePropertyValue;
}
2)After this , it is possible to inject the somePropertyValue into the #Something annotation like this :-
#Something(someProperty = "#{#somePropertyValue}")
public class Foobar {
//...
}
Second solution :-
1) create getter setter in bean :-
#Component
public class config{
#Value("${config.somePropertyValue}")
private String somePropertyValue;
public String getSomePropertyValue() {
return somePropertyValue;
}
public void setSomePropertyValue(String somePropertyValue) {
this.somePropertyValue = somePropertyValue;
}
}
2)After this , it is possible to inject the somePropertyValue into the #Something annotation like this :-
#Something(someProperty = "#{config.somePropertyValue}")
public class Foobar {
//...
}

Try this solution for Java 8
public static void main(String[] args) throws Exception {
final Something oldAnnotation = (Something) Foobar.class.getAnnotations()[0];
System.out.println("oldAnnotation = " + oldAnnotation.someProperty());
Annotation newAnnotation = new Something() {
#Override
public String someProperty() {
return "another value";
}
#Override
public Class<? extends Annotation> annotationType() {
return oldAnnotation.annotationType();
}
};
Method method = Class.class.getDeclaredMethod("annotationData", null);
method.setAccessible(true);
Object annotationData = method.invoke(getClass(), null);
Field declaredAnnotations = annotationData.getClass().getDeclaredField("declaredAnnotations");
declaredAnnotations.setAccessible(true);
Map<Class<? extends Annotation>, Annotation> annotations = (Map<Class<? extends Annotation>, Annotation>) declaredAnnotations.get(annotationData);
annotations.put(Something.class, newAnnotation);
Something modifiedAnnotation = (Something) Foobar.class.getAnnotations()[0];
System.out.println("modifiedAnnotation = " + modifiedAnnotation.someProperty());
}
#Something(someProperty = "some value")
public static class Foobar {
}
#Retention(RetentionPolicy.RUNTIME)
#interface Something {
String someProperty();
}

i am able to access and modify annotaions in this way in jdk1.8,but not sure why has no effect,
try {
Field annotationDataField = myObject.getClass().getClass().getDeclaredField("annotationData");
annotationDataField.setAccessible(true);
Field annotationsField = annotationDataField.get(myObject.getClass()).getClass().getDeclaredField("annotations");
annotationsField.setAccessible(true);
Map<Class<? extends Annotation>, Annotation> annotations = (Map<Class<? extends Annotation>, Annotation>) annotationsField.get(annotationDataField.get(myObject.getClass()));
annotations.put(Something.class, newSomethingValue);
} catch (IllegalArgumentException | IllegalAccessException e) {
e.printStackTrace();
} catch (NoSuchFieldException e) {
e.printStackTrace();
} catch (SecurityException e) {
e.printStackTrace();
}

Annotation attribute values have to be constants - so unless you want to do some serious byte code manipulation it won't be possible. Is there a cleaner way, such as creating a wrapper class with the annotation you desire?

Test custom constraint annotations with JUnit 5

Implementing a custom constraint annotation, like #MySize requires me testing it with unit tests to see if it functions correctly:
public class MySizeTest {
#Test
public void noMinMax() {
Dummy dummy = new Dummy();
// some asserts or so
dummy.setMyField("");
dummy.setMyField(null);
dummy.setMyField("My text");
}
#Test
public void onlyMin() {
// change #MySize to have min: #MySize(min = 1)
... how?
... then test with some setMyField:
Dummy dummy = new Dummy();
// some asserts or so
dummy.setMyField("");
dummy.setMyField(null);
dummy.setMyField("My text");
}
#Test
public void onlyMax() {
// change #MySize to have max: #MySize(max = 50)
...
}
#Test
public void bothMinMax() {
// change #MySize to have min and max: #MySize(min = 1, max = 50)
...
}
private class Dummy {
#MySize()
String myField;
public String getMyField() {
return myField;
}
public void setMyField(String myField) {
this.myField = myField;
}
}
}
I assume this has to be done with reflection, but I have no idea how.

Basicly don't have to use reflection just create a Validator instance and use that for validating.
For examaple:
When the annotation is:
#Target(ElementType.FIELD)
#Retention(RetentionPolicy.RUNTIME)
#Constraint(validatedBy = MyValidator.class)
public #interface MyAnnotation {
String message() default "Invalid value (it must be foo)";
Class<?>[] groups() default {};
Class<? extends Payload>[] payload() default {};
}
and the related validator is:
public class MyValidator implements ConstraintValidator<MyAnnotation, String> {
#Override
public boolean isValid(String s, ConstraintValidatorContext constraintValidatorContext) {
if (null == s) return true;
return "foo".equalsIgnoreCase(s);
}
}
Then the tests sould be like these:
#TestInstance(TestInstance.Lifecycle.PER_CLASS)
public class MyValidatorTest {
private Validator validator;
#BeforeAll
void init() {
ValidatorFactory factory = Validation.buildDefaultValidatorFactory();
validator = factory.getValidator();
}
private static class TestObject {
#MyAnnotation
private String testField;
TestObject() {
this(null);
}
TestObject(String value) {
testField = value;
}
public String getTestField() {
return testField;
}
public void setTestField(String testField) {
this.testField = testField;
}
}
#Test
void shouldValidForNullValue() {
var obj = new TestObject();
var violations = validator.validate(obj); // Set<ConstraintViolation<TestObject>>
Assertions.assertTrue(violations.isEmpty(), String.format("Object should valid, but has %d violations", violations.size()));
}
#Test
void shouldValidForFooValue() {
var obj = new TestObject("foo");
var violations = validator.validate(obj); // Set<ConstraintViolation<TestObject>>
Assertions.assertTrue(violations.isEmpty(), String.format("Object should valid, but has %d violations", violations.size()));
}
#Test
void shouldInvalidForBarValue() {
var obj = new TestObject("bar");
var violations = validator.validate(obj); // Set<ConstraintViolation<TestObject>>
Assertions.assertEquals(1, violations.size());
}
}
Update (2020.05.21.) - Using attributes and AnnotationFactory
Based on comments I've updated my answer.
If you want to test only the validation logic then just create an Annotation instance and call the isValid method which is returns true or false
Hibernate Validator provides AnnotationFactory.create(...) method to make annotaion instance.
After that you can create an instance of your custom validator and call initialize and isValid methods in your test case.
#Target(ElementType.FIELD)
#Retention(RetentionPolicy.RUNTIME)
#Constraint(validatedBy = MyHasAttributesValidator.class)
public #interface MyAnnotationHasAttributes {
String message() default "Invalid value (it must be foo)";
Class<?>[] groups() default {};
Class<? extends Payload>[] payload() default {};
int attributeOne() default 10;
int attributeTwo() default 20;
}
related validator:
public class MyHasAttributesValidator implements ConstraintValidator<MyAnnotationHasAttributes, String> {
private MyAnnotationHasAttributes ann;
#Override
public void initialize(MyAnnotationHasAttributes constraintAnnotation) {
ann = constraintAnnotation;
}
#Override
public boolean isValid(String s, ConstraintValidatorContext constraintValidatorContext) {
if (null == s) return true;
return s.length() >= ann.attributeOne() && s.length() < ann.attributeTwo();
}
}
and the modified test (which has failing assertion):
public class HasAttributeValidatorTest {
private MyAnnotationHasAttributes createAnnotation(Integer one, Integer two) {
final Map<String, Object> attrs = new HashMap<>();
if (null != one) {
attrs.put("attributeOne", one);
}
if (null != two) {
attrs.put("attributeOne", two);
}
var desc = new AnnotationDescriptor.Builder<>(MyAnnotationHasAttributes.class, attrs).build();
return AnnotationFactory.create(desc);
}
#ParameterizedTest
#MethodSource("provideValues")
void testValidator(Integer one, Integer two, String input, boolean expected) {
MyAnnotationHasAttributes ann = createAnnotation(one, two);
MyHasAttributesValidator validator = new MyHasAttributesValidator();
validator.initialize(ann);
var result = validator.isValid(input, null);
Assertions.assertEquals(expected, result, String.format("Validation must be %s but found: %s with params: %d, %d, %s", expected, result, one, two, input));
}
private static Stream<Arguments> provideValues() {
return Stream.of(
Arguments.of(null, null, null, true),
Arguments.of(null, 20, "foo", true),
Arguments.of(null, null, RandomStringUtils.randomAlphabetic(30), false)
);
}
}
Limitations of this solution
Vendor lock
In this case your test using Hibernate Validator which is a specific implementation if the Bean Validation standards. Honestly I don't think it's a huge problem, because Hibernate Validator is the refecerence implementation and the most popular bean validation library. But technically it's a vendor lock.
Cross field validation is unavailable
This soulution works only in one-field situations. If you have e.g a cross-field validator (e.g. password and confirmPassword matching) this example won't fit.
Type independent validation needs more work
Like previously mentioned #Size annotation belongs to several different validator implementations based on type (primitives, collections, string, etc.).
Using this solution you always have to chose the certain validator manually and test it.
Only the isValid method can be tested
In this case you won't be able to test another things just the isValid method. I mean e.g. error message has expected format and parameters or something like this.
In sort, I know creating many different fields with different annotation attributes is boring but I strongly prefer that way because you can test everything you need about your validator.

jdbi BindBean userdefined property of the bean(nested object)

I have a bean class
public class Group{string name;Type type; }
and another bean
public class Type{String name;}
Now, i want to bind group by using jdbi #BindBean
#SqlBatch("INSERT INTO (type_id,name) VALUES((SELECT id FROM type WHERE name=:m.type.name),:m.name)")
#BatchChunkSize(100)
int[] insertRewardGroup(#BindBean ("m") Set<Group> groups);
How can i bind the user defined object's property as member of the bean??

You could implement your own Bind-annotation here. I implemented one that I am adopting for this answer. It will unwrap all Type ones.
I think it could be made fully generic with a little more work.
Your code would look like this (please note that m.type.name changed to m.type):
#SqlBatch("INSERT ... WHERE name=:m.type),:m.name)")
#BatchChunkSize(100)
int[] insertRewardGroup(#BindTypeBean ("m") Set<Group> groups);
This would be the annotation:
#BindingAnnotation(BindTypeBean.SomethingBinderFactory.class)
#Retention(RetentionPolicy.RUNTIME)
#Target({ElementType.PARAMETER})
public #interface BindTypeBean {
String value() default "___jdbi_bare___";
public static class SomethingBinderFactory implements BinderFactory {
public Binder build(Annotation annotation) {
return new Binder<BindTypeBean, Object>() {
public void bind(SQLStatement q, BindTypeBean bind, Object arg) {
final String prefix;
if ("___jdbi_bare___".equals(bind.value())) {
prefix = "";
} else {
prefix = bind.value() + ".";
}
try {
BeanInfo infos = Introspector.getBeanInfo(arg.getClass());
PropertyDescriptor[] props = infos.getPropertyDescriptors();
for (PropertyDescriptor prop : props) {
Method readMethod = prop.getReadMethod();
if (readMethod != null) {
Object r = readMethod.invoke(arg);
Class<?> c = readMethod.getReturnType();
if (prop.getName().equals("type") && r instanceof Type) {
r = ((Type) r).getType();
c = r.getClass();
}
q.dynamicBind(c, prefix + prop.getName(), r);
}
}
} catch (Exception e) {
throw new IllegalStateException("unable to bind bean properties", e);
}
}
};
}
}
}

Doing this in JDBI is not possible , you have to bring out the property and give is a argument.

JSR269 annotation processing getElementsAnnotatedWith() return all the annotated elements each loop and cannot distinguish it belongs to which type

Recently, I have been intersted with the JSR-269 annotation processing and I want to write a lib to eliminate some boilerplate code with it, such as json processing. I really generate the code, then, however, I encounter a fatal problem and spent a lot of time but could not slove it.
The problem is, the RoundEnvironment.getElementsAnnotatedWith() method alway return all the elements annotated with the annotation, I cannot tistinguish which one comes from a specific class. Maybe the problem isn' t clear. I show you the code below.
#Retention(RetentionPolicy.SOURCE)
#Target(ElementType.TYPE)
public #interface JsonObject {
}
#Retention(RetentionPolicy.SOURCE)
#Target(ElementType.FIELD)
public #interface JsonField {
String value() default "";
}
#JsonObject
public class Name {
#JsonField("last") public String last;
#JsonField("first") public String first;
}
#JsonObject
public class User {
#JsonField("age") public int age;
#JsonField("name") public String name;
#JsonField("sex") public boolean sex;
}
the fisrt 2 are the annotation, JsonObject indicates that the annotated type is an JsonObject and JsonField indicates that the annotated field is an json field.
the latter 2 are the sample POJO class that I want to gennerate json-parse code.
in the Processor class, the AbstractProcessor.process(Set<? extends TypeElement> annotations, RoundEnvironment roundEnv) method, when I call roundEnv.roundEnv.getElementsAnnotatedWith(JsonField.class) each type(Name, User) in a loop, the return result is all the json field, in the sample above, the result is ["last", "first", "age", "name", "sex"]. In this situatin, I cannot distinguish which field belongs to which POJO.
May be the words cannot explain what I mean. Here is what I do in the process method.
// the set[Name, User]
Set<? extends Element> jsonObjects = roundEnv.getElementsAnnotatedWith(JsonObject.class);
for (Element jsonObject : jsonObjects) {
// the set[last, first, age, name, sex], **THIS IS THE PROBLEM**
Set<? extends Element> annotatedElements = roundEnv.getElementsAnnotatedWith(JsonField.class);
// other stuff...
}
Feel free to ask me anything I didn' t mention or not clear. Any suggestion is appreciated, Thanks in advance!

You can build a collection of your json object elements by calling the getElementsAnnotatedWith(JsonField.class) method and filering the result based on the annotation of the enclosing element.
Here is a complete example (using runtime annotation processing for simplicity):
#SupportedSourceVersion(SourceVersion.RELEASE_7)
#SupportedAnnotationTypes("*")
public class ElementFilterProcessor extends AbstractProcessor {
#Retention(RetentionPolicy.RUNTIME)
public static #interface JsonObject {}
#Retention(RetentionPolicy.RUNTIME)
public static #interface JsonField { String value(); }
#JsonObject
public class Name {
#JsonField("last") public String last;
#JsonField("first") public String first;
}
#JsonObject
public class User {
#JsonField("age") public int age;
#JsonField("name") public String name;
#JsonField("sex") public boolean sex;
}
public static void main(String[] args) {
JavaCompiler compiler = ToolProvider.getSystemJavaCompiler();
final JavaCompiler.CompilationTask task = compiler.getTask(
null,
null,
null,
null,
Collections.singleton(ElementFilterProcessor.class.getName()),
Collections.EMPTY_SET);
task.setProcessors(Collections.singleton(new ElementFilterProcessor()));
task.call();
}
#Override
public boolean process(
final Set<? extends TypeElement> annotations, final RoundEnvironment roundEnv) {
if (roundEnv.processingOver()) {
return true;
}
final Set<? extends Element> jsonFields =
roundEnv.getElementsAnnotatedWith(JsonField.class);
final Map<Element, List<Element>> jsonObjects = new HashMap<>();
for (final Element element : jsonFields) {
final Element classElement = element.getEnclosingElement();
if (classElement.getAnnotation(JsonObject.class) != null) {
List<Element> list = jsonObjects.get(classElement);
if (list == null) {
list = new ArrayList<>();
jsonObjects.put(classElement, list);
}
list.add(element);
}
}
System.out.println(jsonObjects);
return false;
}
}
Output:
{stackoverflow.annotation.ElementFilterProcessor.User=[age, name, sex], stackoverflow.annotation.ElementFilterProcessor.Name=[last, first]}
I'd also recommend to rather consider using a third-party library for mapping Java objects to JSON. For example, the Jackson processor

Get list of fields with annotation, by using reflection

I create my annotation
public #interface MyAnnotation {
}
I put it on fields in my test object
public class TestObject {
#MyAnnotation
final private Outlook outlook;
#MyAnnotation
final private Temperature temperature;
...
}
Now I want to get list of all fields with MyAnnotation.
for(Field field : TestObject.class.getDeclaredFields())
{
if (field.isAnnotationPresent(MyAnnotation.class))
{
//do action
}
}
But seems like my block do action is never executed, and fields has no annotation as the following code returns 0.
TestObject.class.getDeclaredField("outlook").getAnnotations().length;
Is anyone can help me and tell me what i'm doing wrong?

You need to mark the annotation as being available at runtime. Add the following to your annotation code.
#Retention(RetentionPolicy.RUNTIME)
public #interface MyAnnotation {
}

/**
* #return null safe set
*/
public static Set<Field> findFields(Class<?> classs, Class<? extends Annotation> ann) {
Set<Field> set = new HashSet<>();
Class<?> c = classs;
while (c != null) {
for (Field field : c.getDeclaredFields()) {
if (field.isAnnotationPresent(ann)) {
set.add(field);
}
}
c = c.getSuperclass();
}
return set;
}