This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
We discovered this problem during work on new projects. It is easily reproducible on our projects prototype (<http://projects.netbeans.org/proposals/prototype/index.html>). Start it and expand all nodes of Sample Project, then switch to filesystem tab and mount any new filesystem. All existing filesystems will be removed first and then added. This causes a lot of troubles. The problem is not 100% reproducible, but recently it is happing (if Sample Project nodes were expanded first) almost often. If you are Sun employee (my appologize to non-Sun people) you can download build of the prototype from <http://dkonecny-sun.czech:8080/BuildResults?days=1&branch=prj40_prototype>
*** Issue 27760 has been marked as a duplicate of this issue. ***
*** Issue 27743 has been marked as a duplicate of this issue. ***
Repository.addFileSystem is transparent enough and probably can`t cause such behaviour. This symptom used to be problem of AutomountSupport.
David, can you please reproduce the problem with logging enabled by -Dorg.netbeans.core.AutomountSupport=-1 and attach the output? Thanks for your help.
Created attachment 7637 [details] requested logfile
The problem is in file: core/ui/src/org/netbeans/core/ui/UIUtils.java in method allRoots(...). This method returns new objects for each invocation. But if the folder "Mount" is refreshed (e.g. by adding new filesystem) the method allRoots is called again and returns array of different objects. I am reassigning to Radek since he told me that he plans to reimplement the functionality differently.
Thanx for evaluating the problem. I would like to ask just why the priority was decreased? For projects it is very serious problem because it causes armful of exceptions. It simply breaks everything. I filed it intentially as P2 to get it fixed relatively soon. Please provide your reasoning or increase priority to P2. Thanx.
Because this functionality will be completely reimplemented in DelegatingFileSystem. So, to fix it now seems to us as a waste of a time. But be free to increase priority, if you feel this as a problem.
DelegatingFileSystem is issue 25661, right? Yes, if this will be completely rewritten in near future then it make sense to postpone this fix. I will make this issue dependant on issue 25661 and once 25661 gets implemented this one can be close as WONTFIX. If there will be some complications in issue 25661 we would have to return to this one and fix it. Thanx for explanation.
David, issue 25661 is already fixed, can you try your described scenario again ? If no problem arises, close this as WONTFIX, thanks in advance.
OK, I'm closing this issue. The DelegatingFS should be used and so this problem should not occur. For the moment I'm ignoring filesystems like FTPFS.
.