| [mylyn-integrators] Re: Mylyn Integration with History view |
Mark Phippard wrote:
Right, it won't work with IssueZilla and probably with other connectors except JIRA. The thing is that JIRA task ids usually allow to identify repository right from the id, while IssueZilla and Bugzilla numeric ids aren't unique for repositories.Remote repositories will work if corresponding connector were able to recognize pattern for given task. For example, try to open http://jira-dashboard.tigris.org/svn/jira-dashboard/trunk/org.tigris.jira.ui in the History view and add make sure that http://developer.atlassian.com/jira JIRA repository is registered in Mylyn.I do not understand how that works though. Does it go through every registered provider? I have the three CollabNet connectors installed. In this case, it is the Subclipse repository on tigris and I have the Issuezilla connector installed.
Basically OpenCorrespondingTaskAction.reconcile() iterates trough all connectors and passing comment text to AbstractRepositoryConnector.getTaskIdsFromComment(). Then JiraRepositoryConnector.getTaskIdsFromComment() is using locally cached project keys retrieved from JIRA to construct matching patterns for task keys.
Yes, that is what I meant. Method SubclipseLinkedTaskInfo.init() could be extended to retrieve URL from mapping maintained by Subclipse integration for Mylyn. I was just getting into the technical details where to save such mapping and what kind of UI should be used to manage that.The only option for Subclipse would be to maintain mapping between task repositories and SVN repositories like you suggested (or maybe even per-url mapping). But I am not sure where we can store such mapping. Do you have any ideas on that?I was hoping the Subclipse/Mylyn integration plugin could manage the mapping and it would be a matter of how do we provide the mapping to Mylyn at runtime.
regards, Eugene