Towards User-Level Extensibility of an Ada Library: An Experiment with Cheddar

In this article, we experiment a way to extend an Ada library called Cheddar. Cheddar provides a domain specific language. Programs written with this domain specific language can be interpreted in order to perform real time scheduling analysis of real time systems. By the past, different projects showed that the Cheddar programming language is useful for the modeling of real time schedulers. But these experiments also showed that the interpreter is lacking of efficiency in case of large scheduling simulations. In this article, by designing a Cheddar metamodel, we investigate on how to compile such Cheddar programs in order to extend the Cheddar library. For such a purpose, we use Platypus, a meta CASE Tool based on EXPRESS. For a given Cheddar program and with a meta-model of Cheddar handled by Platypus, we can generate a set of Ada packages. Such Ada packages can be compiled and integrated as builtin schedulers into Cheddar. Then, the efficiency of scheduling simulations can be increased.