Chapter 14. Defining Aspect Instantiation
In Chapter 9, the recipes on advice hinted at an interesting characteristic of how AspectJ implements aspects. When AspectJ was being designed, the decision was made to reuse as much as possible from the existing Java language. This decision meant that even though the nature and intent of aspect orientation was at the forefront of the work on AspectJ, this goal was tempered by an attempt to keep within the constraints of Java so specialized runtime environments would not be needed.
To achieve this goal the designers decided to implement aspects as classes within the post-processed aspect-oriented application. This clever approach meant that 100% of the world’s virtual machines could run AspectJ programs unaltered. This approach is interesting for another reason: since aspects are transformed into regular classes during the weaving process (see Chapter 1 for details), then they must have their own regular object lifecycle. Further, since aspects have their own lifecycle, they must have an associated instantiation policy by which the compiler can decide when a new aspect object is to be created. In this chapter, you see how to declare and control the specific instantiation policies that govern when an aspect object is created within the flow of your application.
14.1. Defining Singleton Aspects
You want to declare that an aspect is to be instantiated as a singleton.
issingleton() explicit aspect instantiation ...