Community
Participate
Working Groups
Merge JIRA connector and existing hyperlink detection capabilities and unify api. See bug#170587 and children. Suggestions that emerged from discussion: - Patterns should be per repository . - Get pattern from connector to detect hyperlinks
Known use cases are: -- Hyperlink detection in task editors (and potentially in any editors, including Java, xml, html, etc) -- Open Correspond Task action from Synchronize and History views
It occurs to me that we may need several detector types. Thinks like "bug# 146964 comment#8" unlikely to appear in commit comments. Also see examples in bug 146964 146964: editor hyperlink parsing problems and suggestions https://bugs.eclipse.org/bugs/show_bug.cgi?id=146964
Update: BugzillaHyperlinkDetector now targets java editor (so bug hyperlinks work in your java source editor). Need to generalize into a 'task' hyperlink detector and generalize to construct 'TaskHyperlinks'.
Also need to investigate use of Eugene's ILinkedTaskInfo and factories to clean this up.
(In reply to comment #3) > Update: BugzillaHyperlinkDetector now targets java editor (so bug hyperlinks > work in your java source editor). Need to generalize into a 'task' hyperlink > detector and generalize to construct 'TaskHyperlinks'. Only want one 'Mylar Tasks' hyperlink contribution so should possibly have a mylar extension point that aggregates connector hyperlink detection.
For Trac we will need a custom hyperlink detector that can not only detect tasks but also detects links to wiki pages.
Rob, consider moving TaskHyperlink and WebHyperlink from Trac to the Mylar tasks plug-in.
(In reply to comment #7) > Rob, consider moving TaskHyperlink and WebHyperlink from Trac to the Mylar tasks > plug-in. Done. What remains here? Perhaps we can we take this further and have a single TaskHyperlinkDetector relying on connector.getTaskIdsFromComment. The only problem would be how to handle WebHyperlinks similarly...
(In reply to comment #8) > What remains here? Perhaps we can we take this further and have a single > TaskHyperlinkDetector relying on connector.getTaskIdsFromComment. The only > problem would be how to handle WebHyperlinks similarly... I'd say that connector.getTaskIdsFromComment() should return collection of *Hyperlinks and not strings. Then we can use it for detecting things like Trac's wiki pages and maybe parent one can also detect urls.
Okay this is what I was thinking so if no one objects I'll continue on this course...
Added AbstractRepositoryConnectoUi.findHyperlinks() which is now used by TaskHyperlinkDetector to detect hyperlinks for all connectors. Removed all connector specific hyperlink detectors and moved connectors specific hyperlink detection code into connectorUi implementation (in case of jira we still call AbstractRepositoryConnector.getTaskIdsFromComment). We can likely now remove unused mylar editor hyperlink extension point code.
