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.
I attempted to create some entity beans from the sample database in pointbase. I picked either the first 2 or first 3 tables in the list for [sample], hit Finish and when the popup window w/ progress bar was up while creating the beans, it got to about 30% and then deadlocked. This is on a 2-CPU Pentium III/600 and having 2 CPU's may be required to get this. I haven't reproduced it, but I haven't tried to. See attached thread dump.
Created attachment 20444 [details] thread dump after deadlock.
Ok, just had it happen a second time just now. Had to kill the IDE same as before as it locks up on shutdown while trying to save MDR. You can save any unsave files though, as far as I can tell, because the CMP generation progress box is a modeless dialog.
Created attachment 20445 [details] another deadlock thread dump, this on at IDE exit
Refined steps: 1. Start IDE. 2. Create empty EJB module project 3. Start pointbase 4. Invoke entity beans from database wizard 5. Choose pointbase\sample 6. Pick the first table only (though this happens with 2-3 and probably more tables selected as well) 7. Hit <ok, finish, whatever> to complete the wizard. When the "Generating CMP's" progress dialog appears, the deadlock happens when it's at about 30%. If it gets to 50%, the deadlock will probably not occur in that attempt.
I worked with this wizard really a lot, but I am not able to reproduce this behaviour. Is QE able to reproduce?
I was trying to reproduce the issue for a hour but without success. I performed above mentioned steps several times. CMP beans were generated everytime. I tested on the configuration: machine: 2-CPU P II jdk: 1.5.0_01 NB build: 20050309
Peter, can you try to reproduce. I expect this bug to be fixed, lots of changes were done in this wizard (including switch from org.openide.src to JMI)
I also did not encountered this problem - the code that instantiate beans in wizards and holds locks is rewriten now as Martin pointed. This is very likely not a problem now.
I will find some cycles to retest this. My guess is that if the creation code has been overhauled, the bug most likely no longer exists (I hope it didn't just move).
*** Issue 56413 has been marked as a duplicate of this issue. ***
http://www.netbeans.org/issues/show_bug.cgi?id=56413 has the same symptoms... We need JMI gurus to take a look at the 2 thread dumps.
and http://www.netbeans.org/issues/show_bug.cgi?id=55439 seems to be related as well, another p2 Maybe not, but worth taking a look...
The first deadlock was caused by a bug in javacore that was fixed by the following commit to the ResourceImpl class: /cvs/java/javacore/src/org/netbeans/modules/javacore/jmiimpl/javamodel/ResourceImpl.java,v revision 1.68 date: 2005/02/24 17:52:21; author: mmatula; state: Exp; lines: +21 -9 Added transaction to prevent deadlocks I am going to investigate other threaddumps.
The second thread dump shows the same bug, so this issue is fixed. The other two issues that Ludo mentioned are unfortunatelly not related to this one.
v.