Bug 603962
Summary: | [abrt] crash in java-1.6.0-openjdk-devel-1:1.6.0.0-39.b18.fc13: Process /usr/lib/jvm/java-1.6.0-openjdk-1.6.0.0/bin/java was killed by signal 6 (SIGABRT) | ||||||||
---|---|---|---|---|---|---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Dan Irwin <rummymobile> | ||||||
Component: | java-1.6.0-openjdk | Assignee: | Omair Majid <omajid> | ||||||
Status: | CLOSED DUPLICATE | QA Contact: | Fedora Extras Quality Assurance <extras-qa> | ||||||
Severity: | medium | Docs Contact: | |||||||
Priority: | low | ||||||||
Version: | 13 | CC: | aph, dbhole, dlila, jon.vanalten, langel, lkundrak, mjw, mmatejov, ptisnovs | ||||||
Target Milestone: | --- | ||||||||
Target Release: | --- | ||||||||
Hardware: | i686 | ||||||||
OS: | Linux | ||||||||
Whiteboard: | abrt_hash:7ddd0e44704d43c291f4717b4672521070587adf | ||||||||
Fixed In Version: | Doc Type: | Bug Fix | |||||||
Doc Text: | Story Points: | --- | |||||||
Clone Of: | Environment: | ||||||||
Last Closed: | 2010-07-15 18:09:31 UTC | Type: | --- | ||||||
Regression: | --- | Mount Type: | --- | ||||||
Documentation: | --- | CRM: | |||||||
Verified Versions: | Category: | --- | |||||||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||||
Cloudforms Team: | --- | Target Upstream Version: | |||||||
Embargoed: | |||||||||
Attachments: |
|
Description
Dan Irwin
2010-06-15 02:46:40 UTC
Created attachment 424014 [details]
File: backtrace
I was able to reproduce this. I tried to reproduce it twice by resizing the properties window, and I was successful. Then I tried 4 times to reproduce it by resizing the main netbeans window, without any projects open. I succeeded 2 of those times, and I had to move the mouse faster and for longer than in the cases where I was resizing the properties window. On the 2 other times I couldn't reproduce it. My arm got tired, so I stopped trying. I could not reproduce the bug with any small, trivial swing programs (though, I didn't try very much). So I thought that this might be related to how quickly the window can be repainted relative to how many repaint requests there are. So I ran two instances of a while(1) program to load the CPU as much as possible, but even in this case I was not able to reproduce the bug by resizing the main netbeans window. I'm able, like Denis, to reproduce this issue. I also tried to change appearance of window borders and title bars using Gnome configuration tools, but it does not prevents crashes of Netbeans & JVM. It is also possible to reproduce this bug using OpenJDK VisualVM tool which is installed with OpenJDK itself (so you don't have to install Netbeans). Because VisualVM's window contains fewer widgets than Netbeans Project Window, it takes longer to crash JVM by resizing it's window. Tested on: Fedora 13 i686, kernel 2.6.34-20 IcedTea6 1.8 (fedora-39.b18.fc13-i386) latest X.org installed Possible clue: I reproduce this on my F12 box (as expected from previous comments). But, I have tried this with upstream openjdk7 and after spending several minutes resizing Properties window I get no crash. I also notice that in general the display updates are snappier. For example if I shrink the Properties window the initially grey background is very quickly replaced with the display of the main window behind it, but with the icedtea6 installed on my system there is much more lag. I do, however, still get the crash using builds from IcedTea6 HEAD or upstream openjdk6. I tried to reproduce this a few times on a distribution of F13 using icedtea6 and running xfce instead of gnome, and no matter how long or how quickly I resized, I couldn't get it to crash. Created attachment 426903 [details]
a tiny reproducer
This is a tiny java program that can reproduce the issue. Simply resize the window fast enough (it may take a few tries) to get the crash.
A few observations:
1. The crash only seems to happen with the gtk look and feel. Using swing's cross platform look and feel, the problem disappears
2. The NOISY_AWT environment variable can be used to make the program print out a stack trace as soon as awt sees an x error (the problem can be observed as soon as the window is made large enough):
NOISY_AWT=abort java ResizeCrasher
3. The problem does not happen under a virtual machine. This program runs fine under gnome and xfce (under kvm).
*** This bug has been marked as a duplicate of bug 551835 *** |