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.
Probably not reproducible. Occurred while editing. Data loss.
Created attachment 12429 [details] Thread dump
Current dev sources.
Happened to me again. Also while editing text. Again, data loss.
too bad, Petr Nejedly is on vacation until the end of Dec. Yarda, please investigate. Mila: can this be related to the threading changes you made in openide/text and editor recently? Thanks
Created attachment 12460 [details] Fixes ordering of locks
I tried to simulate the deadlock but I have failed as there is not a way how to hook into the code and one would have to rely on a luck. Anyway I think I understand the problem and I have a fix. The cause of the deadlock is a code that under private lock manipulates with private HashMap, but this triggers DocumentLine.equals which asks for rendering lock. I have not found a way to prevent the DocumentLine.equals to call outside of the synchronized block, so I decided to order locks: before the private lock is obtained, the DocumentLine.Set obtains rendering lock first, this way any call to DocumentLine.equals has to successfully succeed, as the lock is already being held. Please review.
Created attachment 12461 [details] Happened to me again, new thread dump
With or without the patch? From what I see in DefaultRP, without. As you seem to be able to reproduce this problem really often, you might consider applying the patch to try to run to the same problems with it.
I think Yarda's patch should fix the problem. Unless I have overlooked anything the two threads are counter-locking on doc-read-lock and Line.Set.lines. The attached fix should enter the monitors in the same order.
DocumentLine 1.52
OK, I will let you know if it reappears (I am running with the patch now). Thanks for the quick patch.
verifying, Jesse is silence from 2003-12-08 and so I belive it's gone