Community
Participate
Working Groups
I have a simple extension point in one plug-in, it is contributed to by that same plug-in (one extension entry) and one other plug-in (two extension entries). To demonstrate this logic to a class I added features for each plug-in and installed the set on my workbench instance. I can dynamically disable a the feature that references the plug-in with only a set of extensions and my extension point logic sees one extension entry; its own. If I enable the feature my logic sees three entries; its own and two from the enabled plug-in. When I look for this same content in the plug-in registry view I don't get the same end result. If I go to the plug-in registry view in the beginning (one plug-in with extension point and extension active) I see one extension entry under the extension point node. If I enable the feature to add to the number of extensions found, my logic sees them, but they are not visible in the plug-in registry view; only the one extension is shown under the extension point node for the owning plug-in. A refresh does not change this at all. If I then disable the feature with the plug-in that adds two extensions to the extension point, the extension point node in the plug-in registry view is empty; not even the one extension added by that same plug-in is visible under the extension point node. Refresh does not do anything, but if the view is closed/opened it then displays the correct content. Seems like the plug-in registry view, given that it supports IRegistryChangeListener, should be reacting to changes that occur in the registry either automatically or worst-case, with a refresh.
This may be related to or judged as a duplicate of bug 81322. Note, did just try this again on 3.1, and while I can't dynamically uninstall my feature (that may have been a bug in 3.0 that I could), a dynamic add still shows that the plugin registry view is out of date. I have to close the view to get it to show current content.
we are beginning the process of migrating the registry view to be more dynamically aware. will resolve this bug as dup of bug 81322, as we attempt to solve all refresh problems in that bug. *** This bug has been marked as a duplicate of 81322 ***