dc.contributor.advisor |
|
|
dc.contributor.author |
Wijerthna, Hasan Nimesha |
|
dc.date.accessioned |
2019-03-04T11:18:28Z |
|
dc.date.available |
2019-03-04T11:18:28Z |
|
dc.date.issued |
2018 |
|
dc.identifier.citation |
Wijerthna, H. N. (2018) Commit analyzing tool for Git version control system. BSc. Dissertation. Informatics Institute of Technology |
en_US |
dc.identifier.other |
2014198 |
|
dc.identifier.uri |
http://dlib.iit.ac.lk/xmlui/handle/123456789/164 |
|
dc.description.abstract |
Git is a distributed version control system allows each developer in a software project to
have their local copy of the project that they have been working on. The transparent
mechanism behind the version control system is to associate everyone using pull requests
and commit messages which also can be reviewed later. This vast amount of activity could
relate to implement research models can be derived for better enable productivity and
communication amongst Git developers.
When developers are involved in large repositories with many edits throughout the project
it is crucial to commit with a relevant commit message to ensure every member in the
project to aware about the associated changes of the project. Since these commits make
an informed overall decision about additional modification. The existing system allows
developers to filter commits by the commit message and manually filter those code
changes. If this procedure doing in projects with higher number of commits this may not
be efficient and regardless the only way to identify additions and deletions of files,
libraries and plugins is to manually checking the dependency installation files which not
be efficient in large collection of files to check these changes manually. Most importantly
this procedure of maintaining the code base varies from developer to developer.
In large repositories with many commits throughout the repository, Commit Analyzer was
designed and implemented to be able to flag the commits and pull requests which are
crucial to ensure everyone to be aware of the changes associated with a commit and can
make an informed decision about whether to merge or ask for additional modifications in
major milestones such as in a software release. So this model retrieves crucial details of
source code changes and provides a summarized report in a given period of time.
System accuracy was thoroughly tested based on the rules while accuracy was tested
based on the output values by Commit Analyzer. The usage of Paragraph vector classifier
makes Commit Analyzer perform well. Commit Analyzer produces acceptable results and
is hence justified. The tested results attested that the analysis, design, implementation and
documentation have been carried out in an efficient manner. |
en_US |
dc.subject |
Distributed Artificial Intelligence |
en_US |
dc.subject |
Paragraph vector classifying |
en_US |
dc.title |
Commit analyzing tool for Git version control system |
en_US |