Embedded-Software-Automated-Testing-Oriented Method And System Based On SCC50 DTU

Software testing is an important means to ensure the quality,reliability and security of software products.The manual testing exposed a lot of defects as the scale of the software became large,for example,it consumed lots of manpower and time,but with low efficiency,and it is difficult to carry out regression testing.Compared with manual testing,automatic testing has many advantages,for instance,it costs short time,avoid the human misjudgment and be easy to carry out regression testing.Automating embedded software testing becomes an inevitable trend,and it has emerged as the hotspot in the embedded system research field in recent years.This paper studies the characteristics of embedded software development,the strategy of automatic testing and so on.A software automatic testing system based on an embedded software called SCC50-DTU(Data Transfer Unit)was developed.This system aims to reduce the tester's workload,shorten the development cycle of software under test and improve the testing efficiency.This system uses cross testing strategy,supports unit testing and regression testing;use a database to store all kinds of test data to facilitate the management to test data;it generates the intuitive and detailed testing report of test suite and supports the tester to customize various testing tasks.The effect of the system's use proved that it has good scalability and reliability.This paper used an improved test case prioritization in the regression testing,which algorithmically combines white-box and black-box testing.The prioritization adds the function coverage information to a prioritization which is oriented black-box testing,obtains the function diagram of the source code through static analysis and the actual execution trace of the test case by code instrumentation.Then it calculates a value called "errcode" based on the trace and uses the errcode value for another factor for prioritization to improve the efficiency of defaults detection.This paper designed an experiment to make a comparison of this algorithm with others.The experiment result showed that it is more efficient on the behalf of faults detection.