I have a requirement to create a custom Annotation which when applied over a method checks and validates the input parameters (Primitive and non-primitive) against various checks. And if checks fail should return a error message directly.
While searching I have come across use of AbstractProcessor and ConstraintValidator when we create a custom annotation. Being new to creating custom annotation, I would like to understand how to go ahead implementing my problem statement.
First of all, you need to understand that you're talking about two different things.
TL;DR: compile-time vs run-time - you need run-time
An Annotation processor is an abstract term to identify a process which runs during compile time, and which is able to scan the source files, the ones which have a .java extension.
The Annotation processor implementation might throw warnings, or even errors to stop the compilation task, based on arbitrary checks.
Example: "I noticed someone passed a null value to my method, it's a error! Stop everything!"
The starting point for implementing an Annotation processor is the AbstractProcessor base class, or the Processor interface (documentation).
You'd also need, unlike the ConstraintValidator implementation, to register it explicitly via a processor metadata file, which must be located under a standard directory inside the produced JAR artifact.
META-INF/services/javax.annotation.processing.Processor
What is needed to create an Annotation processor is already included in the standard JDK. You don't need external dependencies.
On the other hand, a ConstraintValidator identifies an interface which is shipped with the Validation API module, under the package javax.validation. There are different implementations of this API, but the most used one is Hibernate Validator (documentation).
The validation which this interface provides are verified at runtime.
Unlike the Annotation processor, the Validation API implementation must be provided manually, e.g.
<dependency>
<groupId>org.hibernate.validator</groupId>
<artifactId>hibernate-validator</artifactId>
<version>6.0.15.Final</version>
</dependency>
You wrote
when applied over a method checks and validates the input parameters (Primitive and non-primitive) against various checks
It seems you need to write run-time checks, which means the Validation API is the road to take.
The use cases of AbstractProcessor and ConstraintValidator are quite different. Let's take a dive and explore both.
The AbstractProcessor is
a convenient superclass for most concrete annotation processors
So what are annotation processors and what is their use case? If you want to generate new source code at compile time, based on annotations, you can use annotation processors. They are provided as part of the compilation process and invoked through the Java compiler. An example of this is a static metamodel geneartors for JPA.
A ConstraintValidator can be used to ensure, at runtime, that attributes, parameters and alike satisfy certain constraint. One of the most popular constraints is @NotNull. At runtime, some piece of code, a constraint validator, checks that all fields/parameters annotated with @NotNull actually are not null. If you want to write your own constraint, you need to write an annotatation, as well as a processor to satisfy your constraint. This tutorial on dzone.com outlines the process of creating a custom annotation and corresponding ConstraintValidator.
