Methods, Part 1

7/21/2007 3:26:28 PM

Methods, Part 1

Method handling is difficult in static analysis. I will describe in a few parts some of the issues that I have had to deal with.

I demonstrated in earlier posts my handling of recursive functions in Loops, Part 1. In this post, I will show how I elegantly handle virtual functions. Below is the code for three classes, a base class Animal and two derived classes, Cat and Dog. Animal is an abstract base class, so no implementation actually exists for one of its member function.

Each animal can make a noise and has a type.

To test the three Animal classes, a wrote the following test method below. We have three variables of type Animal, "spot," "polly," and "tabby." We know Spot is a dog through an Assert statement. We know Tabby is a cat because we just created one from scratch. However, Polly is an unknown and sounds as if it may be a parrot, a class that we actually haven't account for; or, we could be mistaken, and it could actually be a dog or cat. 

So, I run the code through NStatic and extract out all the local variables and its member variables after the scan. Here's what the resulting Locals window looks like.

Whether we assign a variable or assume it, we know through the variables of the form "isXaY" that Spot is a dog and not a cat, and Tabby is a cat and not a dog. We also know that Spot says "Bark" and Tabby saw "Meow." We can also tell each animal's type from Animal's virtual property, Type. However, the Polly variable is mysterious. Thus, the local window has to represent all information about Polly's nature and behavior through conditional expressions and possibly generated values (in the case of Animal.Say(), which has no implementation).







Net Undocumented is a blog about the internals of .NET including Xamarin implementations. Other topics include managed and web languages (C#, C++, Javascript), computer science theory, software engineering and software entrepreneurship.

Social Media