C编译器自动测试仿真研究

该文涉及的C编译器仿真自动测试的主要思想是在自动生成c编译器测试用例的设计中引入随机的思想和算法。一方面，用“任何测试点都可能测到”的思想代替“每个测试点都必须测到”的思想，从而在测试用例的数量达到一定规模时，在满足应用的条件下实现对编译器产品的正确性评估；另一方面，完全排除了人为因素对测试的影响。其另一关键思想是引入一个实践证明的成熟编译器作为参考编译器，让被测编译器和参考编译器在同样环境下运行同一组测试用例。然后对比运行结果，如有不同，即证明被测编译器有错误，这种模式，有效排除了预先给定运行结果的测试模式中人为因素对测试结果的影响。

Researching for Automated Testing of C Compiler

It is the main idea of this paper that to design a random algorithm to generate the test case at random. The random algorithm resolved the problem of the test case generated automaticaly covering with all the point of testing instead of the method of enumerating all things, on the other hand, it prevent the disadvantage brought by the subjectivity of the engineers from the testing. In order to verify the result of testing, this paper import a reference compiler which has been used maturely, and made the reference compiler run the test case at the same environment with the object compiler. Then, it can conduct the testing conclusion through comparing the two sets of result got from the object compiler and the reference compiler.