Hey Guys,
I'm one of the contributors to this new plugin and I'd like to try and explain why we chose to file for hosting instead of extending the Result Publisher plugin.
First off, let me give you a bit of background info: We've been having trouble with mapping test fails to JIRA issues for over a year now and we've been looking into ways of fixing this in order to avoid the overhead of manually querying and searching for related builds/fails/issues. While the Test Result Publisher did meet some of our needs, it had major drawbacks:
- No relation visible in Jenkins to corresponding Issues
- No possibility of configuring the template for the newly created Issues
- High risk of flooding JIRA instance with many issues (we have roughly 3000 tests, which may fail randomly)
The main focus of our plugin is to offer the possibility of linking test fails to JIRA issues, be they existing or newly created. During its development we've realized that there are some other issues with the other plugins, like security and the connectors used (if any), so development just kept going. The main features the plugin currently has are:
- Ability to link existing JIRA issues with tests fails
- Ability to create new JIRA issues for new test fails
- Visual representation of the link in the Publisher page, with issue number/link and status
- Link persistency between builds
- Configuration of Issue template on a job level (including labels, components etc.)
- Thread-safety
- Uses the Atlassian JIRA rest Java library
- Supports multi-configuration projects
We do realize that some people would require the functionality that JIRA Test Result Publisher has, but there would be two options for that:
- use both plugins concurrently
- add the functionality of the Result Publisher plugin to our own
I, for one, would suggest adding the functionality to our infrastructure, because that would take less time to achieve and we have a more robust infrastructure.
All in all, while we are advocates of reusing things and contributing to existing things, this current situation looks like an overly complicated approach to grafting, by tying the whole tree to a branch, rather than tying the branch to the tree.
Regards,
Catalin