Asynchronous Collaborative Programming Semantics-based Version Of The Merger

Semantic-based program versions merging has a vast and an important prospect in applying in the asynchronous collaborative programming, which can automatically merge the code by multiple developers simultaneously modified, and offers the base for final amendments. It sufficiently supports the asynchronous collaborative activity, raising the working efficiency of co-operators. In the environment of asynchronous collaborative programming, the co-operators often construct many versions, of which each is a revision or supplement to the original version. As a result, how to merge so many versions into one becomes a very important research topic.This paper firstly introduces the asynchronous collaborative programming technique, program initialization and tag create on technique, which are the base of program differencing analysis.Then it discusses the available differencing analysis and merging technique. In the process of merging the different versions into one, the chief technique is analysis on the difference among the versions. So the paper emphasizes the analysis mechanism about semantic-based difference: the construction of program dependence graph, the partitioning of sequence-congruence, the finding of affected points, and so on.Through the strategy of program transformation, the algorithm of quick finding the affected points, etc, the result of the sequence-congruence can be enhanced, hence a semantic-based accommodated program difference integration (ADI)algorithm is presented. The ADI uses two kinds of program slices------forward sliceand backward slice at the same time which can avoid occurrence of pseudo-conflict phenomenon, effectively solving the problem that the conflict is reported but in fact there is no conflict among the programs in merge algorithm.Finally, according to the research of semantic-based program integration, a semantic-based prototype system is designed and implemented for merging of the versions. All algorithm and schemes presented in this paper have been implemented and verified by the system .Experiment's results show that these algorithms proposed in this paper can effective and feasible from implementation viewpoint.