Biblio

Dynamic tainting is an important part of modern software engineering research. State-of-the-art tools for debugging, bug detection and program analysis make use of this technique. Nonetheless, the research area based on dynamic tainting still has open questions, among others the automatic generation of program inputs. My proposed work concentrates on the use of dynamic tainting for test case generation. The goal is the generation of complex and valid test inputs from scratch. Therefore, I use byte level taint information enhanced with additional static and dynamic program analysis. This information is used in an evolutionary algorithm to create new offsprings and mutations. Concretely, instead of crossing and mutating the whole input randomly, taint information can be used to define which parts of the input have to be mutated. Furthermore, the taint information may also be used to define evolutionary operators. Eventually, the evolutionary algorithm is able to generate valid inputs for a program. Such inputs can be used together with the taint information for further program analysis, e.g. the generation of input grammars.

Knowing which part of a program processes which parts of an input can reveal the structure of the input as well as the structure of the program. In a URL textlesspretextgreaterhttp://www.example.com/path/textless/pretextgreater, for instance, the protocol textlesspretextgreaterhttptextless/pretextgreater, the host textlesspretextgreaterwww.example.comtextless/pretextgreater, and the path textlesspretextgreaterpathtextless/pretextgreater would be handled by different functions and stored in different variables. Given a set of sample inputs, we use dynamic tainting to trace the data flow of each input character, and aggregate those input fragments that would be handled by the same function into lexical and syntactical entities. The result is a context-free grammar that reflects valid input structure. In its evaluation, our AUTOGRAM prototype automatically produced readable and structurally accurate grammars for inputs like URLs, spreadsheets or configuration files. The resulting grammars not only allow simple reverse engineering of input formats, but can also directly serve as input for test generators.