Community
Participate
Working Groups
Build 20040420 As soon as a PRE_DELETE event is notified on a project, it is removed from the JavaModel. However, if for some reason, the actual deletion never occurs (e.g. file in it is being used by some process), the project is still present in navigator, but gone from package explorer.
2 issues behind: - model update should be deferred until actual DELETE is received. Maybe some issues about list of new/old projects for delta computation. - indexing may be busy on this project. If we don't discard associated jobs, it may cause the removal to fail (we could keep handles to files in it); but if we remove them, and the project is not actually removed, then our index may be inconsistent. Suggestion: we discard all its indexes eagerly on PRE_DELETE, and if not actually deleted, then indexes will be recreated lazily from scratch.
Deferred since we had this bug since day-1 and never got hit by it.
Reopening
Deferring post 3.1
Our source control plugin forbids the deletion of entire projects... however we see this precise behavoir in the following scenario: If you delete a source-controlled project and select "also delete the source..." our deleteProject method rejects the deletion. However, suppose you have two or three other projects that are not under source control... say on the "C" drive in the workspace. If you select one of the deleteable projects and say, accidentally, multiply select the source controlled one, nothing unusual happens. BUT if you multiply select two or more deleteable projects then inadvertently also select a source-controlled project, you get the warning that you can't delete a source controlled project, and the selected deletable projects do in fact get deleted, but the non-deletable source-controlled project dissapears from the package explorer, but it is clearly still in the workspace (as evidenced by the fact that it's in the list when you do a team->share or that it does not show up if you attempt a re-import because you think it's gone...) This behavior is present in Eclipse 3.2M6
John this is quite late in the 3.2 cycle, and this code is risky to change. We will consider this bug post 3.2
I understand perfectly... note below says it's been there for a long, long, time... so I thought that a possible use case for reproduction might be useful.
As of now 'LATER' and 'REMIND' resolutions are no longer supported. Please reopen this bug if it is still valid for you.