Embedded development: skills and tricks -- seven skills to improve the quality of embedded software code

​　　 For developers , Nothing is worse than trying to maintain and update code of such poor quality . For many development teams , Software code quality is not required . To improve code quality , Here are a few tips that embedded developers can follow , And integrate it into your own process , This will help improve their code quality .

　　 skill 1—— Follow coding standards

　　 Developers interested in improving code quality should start by examining industry standards . Developers can use many industry standards to improve the way they write code , Include :

　　l MISRA-C

　　l Cert-C

　　l IEC 12207

　　l IEC 62304

　　 These standards establish a framework for developing high-quality software . Anyone who wants to improve software quality should review what has been done , And start implementing industry recommendations .

　　 skill 2—— Analyze and maintain software complexity

　　 It has been proved many times , Minimizing software complexity also minimizes error rates . The more complex the software is , The greater the change , Not only will there be software bug, And the maintenance of the system will also increase bug. Developers should monitor their function complexity by using cyclomatic complexity .

　　 skill 3—— Conduct regular code reviews

　　 Code review has proved to be the most effective way to reduce software errors . Integrating peer reviews into the development cycle is crucial to ensure that software quality remains at a high level . In order to have an effective code review process , Embedded developers should follow several key factors :

　　 Limit code reviews to 60 Within a

　　 Don't comment more than 400 individual LOC

　　 Schedule a code review at least once a week ( If you generate more than per week 400 individual LOC, More )

　　 skill 4—— Requirements must be traceable

　　 If a requirement cannot be tested and verified in some way , It's better to be automated , Then it is not a real demand . Even before the development starts , Requirements should be broken down into traceable test cases , These use cases can be used later to verify that all requirements have been met . Of course , Needs change , Will use different design methods , But there needs to be a way to track software features and functions , Trace back to the requirements that produced them , And prove that they work as expected .

　　

　　 skill 5—— Compile without errors and warnings

　　 Many teams and developers who ignore compiler warnings , Sometimes there are only oneortwo warnings that developers need to dig deep , Or trivial , So they were ignored , Just hang there . If the compiler does not trust the code and gives a warning , Then developers should treat it as a compiler error and resolve the warning . Warnings are the way the compiler says it doesn't like something . honestly ,C Compilers like everything , So if it has problems with something , So as embedded developers, we should do the same . Code quality means no errors and warnings .

　　 skill 6– Static analysis

　　 Static code analyzer is one of the fastest ways to find potential problems in embedded software . Static code analyzer can find syntax problems 、 Potential glitches , And identify the non portable structure in the code , The static analyzer can even check whether the coding standard is followed , Integrating static analyzers into a continuous build process is a good way to improve code quality and minimize software errors .

　　 skill 7—— All code must be thoroughly tested

　　 Many small and medium-sized enterprises release products without completely testing the software . These companies tend to carry out high-level spot checks , To see if the system works as they expect , Then send them to the wild . When most readers learn that the products of these companies encounter problems in this field , May not be surprised . As an embedded developer , We can't think that our code is working properly just because we don't see the improper behavior of the system . Our code needs to be fully tested ! Comprehensive testing means :

　　l High level test cases have been verified

　　l Unit tests have been performed

　　l Integration tests have been performed

　　l Code coverage testing has been completed , To verify that each line of code is executed during the test

　　l System tracking analysis , To verify the time 、 Performance and system behavior are correct , It's not just right for users .

　　 If a line of code is not executed during the test , There is no way to know whether this line of code contains errors that will occur when users use the system one day .

　　 Conclusion

　　 Improving code quality is not necessarily a major 、 Expensive tasks . These suggestions are easy to implement , It can even be completed in a few weeks or months . The end result will be higher quality software , Provide a better user experience , At the same time, reduce the cost of embedded development .