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 was working with a project I had available in the Shared Files view in an open collaboration session. The remote party ended his session. My IDE got frozen. See the attached thread dump.
Created attachment 25491 [details] Thread Dump
Was there any exception? From the thread dump, you see that AWT waits for one of the RPs to finish ("Java Children Provider"), but all 4 RPs are waiting for access to the ExclusiveMutex. I can't find a single place where ExclusiveMutex should be held, so either I overlooked something, or the thread locking EM died w/o proper unlock in finally{} block. In either case, it looks like collab disconnect have just uncovered a bug somewhere else (editor, javacore, openide/text). Marking as INCOMPLETE for now and reassigning to javacore.
> Was there any exception? No, it simply hang.
Is this still INCOMPLETE? What do you need to know more?
We need to know steps how to reproduce and complete message.log.
Hmm, I tried reproducing, but this time the collab handled the other party's disconnection gracefully. Sorry, no more info.
I did encounter deadlocks of the IDE a few times while using the Collab module - and I have to admit that it is rather random (it once got stuck while I was sitting there moving my cursor in the conversation window!) Hence the issue report might not be describing the problem precisely, and that occurance of such problems should be rare, lowering the priority of this issue.
Alexlamsl, please, don't change priority of bugs which you did not report and which are not assigned to you. The issue owner should make the evaluation. Thanks! Anyway I believe that deadlock issue cannot be P4. It can either be changed to RANDOM, WORKSFORME or FIXED. Putting Marian from QE on CC, who could help us reproduce the problem or verify that it does not appear anymore. Mariane, please, have a look at the problem and try to reproduce it. I was not successfull anymore with reproducing.
*** Issue 70575 has been marked as a duplicate of this issue. ***
It seems that I have managed to reproduce the problem with unfinished transaction, it simply happened after opening nb project and its required projects and the classpath scanning never ended. Progress bar of the scanning was painted on start and no action that relies on scanning could be finished. When attempting to exit IDE it deadlocked. Attaching thread dumps from the living IDE when scanning was running forever and from deadlocked IDE during exit.
Created attachment 28130 [details] thread dumps
It looks like some transactions are sometimes not ended properly. We need to investigate this issue for 5.0.
*** Issue 70902 has been marked as a duplicate of this issue. ***
I have probably another thread dump showing the same problem, deadlock happened during IDE shutdown.
Created attachment 28158 [details] thread dump
MDR transactions are not correctly ended in sporadic cases. After serious error in application (StackOverflowError, OutOfMemoryError) current MDR transaction is sometimes not finished. This unfinished transaction can end in deadlock. The only reproducible test case happens after this StackOverflowError http://www.netbeans.org/issues/show_bug.cgi?id=70774. We already fixed this case, but we are not sure, that there are not any other cases, when unfinished MDR transaction may lead to deadlock. This is reason, why we want to keep this issue open for now.
Should be fixed by thos commit: http://mdr.netbeans.org/servlets/ReadMsg?list=cvs&msgNo=1578
This has become obsolete with interduction of Retouche. Marking as Verified. --- NetBeans IDE Dev (Build 070214) 1.6.0; Java HotSpot(TM) Server VM 1.6.0-b105 Linux version 2.6.12-1.1390_FC4smp running on i386 en_US (nb); UTF-8
Reorganization of java component