There are six general areas of SCM deployment, and some best practices within each of those areas. The first, second, third and fourth area and there practices are already discussed.
Builds
Builds are necessary to construct software from source files. The high level practices involved with this workspace are :
- In a build, the only inputs that are needed are the source files and the tools that are used. There is no place for procedures or yellow sticks. If setup procedures are there, automate them in scripts. If manual setup is there, document them in build instructions.
- There are chances that the input list is incomplete if the software cannot be build from the same inputs. So, the input list needs to be checked.
- When you are organizing original source files in a directory, you need to ensure that already built objects are kept away and do not contaminate the source files. Built objects (those files that get created during the building process) should be located in a different directory, away from the original source files (which are the files that have been created through a tool such as a code editor, or through notepad, or through code generation tools).
- Developers, test engineers, and release engineers should all use the same and easily available build tools.
- Build should be done often as end-to-end builds with regression testing (“sanity” builds) reveal integration problems introduced by check-ins and they produce link libraries and other built objects that can be used by developers.
- Build outputs and logs, including source file versions, tools and OS version info, compiler outputs, intermediate files, built objects, and test results should be kept for future reference.
No comments:
Post a Comment