abrt version: 1.1.14 architecture: i686 cmdline: /usr/bin/python -Es /usr/bin/sealert -s comment: See above. component: setroubleshoot executable: /usr/bin/sealert kernel: 2.6.35.10-74.fc14.i686 package: setroubleshoot-server-3.0.15-1.fc14 reason: browser.py:648:on_delete_button_clicked:IndexError: list index out of range release: Fedora release 14 (Laughlin) time: 1293892828 uid: 500 backtrace ----- browser.py:648:on_delete_button_clicked:IndexError: list index out of range Traceback (most recent call last): File "/usr/lib/python2.7/site-packages/setroubleshoot/browser.py", line 648, in on_delete_button_clicked self.database.delete_signature(self.alert_list[self.current_alert].sig) IndexError: list index out of range Local variables in innermost frame: widget: <gtk.Button object at 0x9cb966c (GtkButton at 0x85f81e8)> self: <setroubleshoot.browser.BrowserApplet instance at 0x88c3e2c> How to reproduce ----- 1. I was troubleshooting SELinux alerts related to Redmine installation. 2. The SE Troubleshooter was open and I saw a new alert notification, but nothing was displayed. 3. I closed SE Troubleshooter, clicked on alert icon and a blank alert was displayed. I tried to delete that blank alert.
Created attachment 471333 [details] File: backtrace
Had you told the troubleshooter to ignore alerts?
I do not remember for sure if I clicked on Ignore button before the Troubleshooter crashed. I was getting lots of alerts every time I started Redmine application. Most likely, I have deleted all the alerts received, but not ignored them.
If you run sealert -s in a terminal window, are you seeing a back trace before this? Could you attach your /var/lib/setroubleshoot/setroubleshoot_database.xml
Hi Daniel, I'm not sure I understood your question about the sealert -s. If I run that command in a terminal, the SELinux Alert Browser opens. Currently there is no alert pending, so I do not see any alert notification in the browser. This is the content of /var/lib/setroubleshoot/setroubleshoot_database.xml: <?xml version="1.0" encoding="utf-8"?> <sigs version="3.0"> <users version="1.0"> <user_list> <user username="pires" version="1.0"> <email_alert>False</email_alert> </user> <user username="root" version="1.0"> <email_alert>False</email_alert> </user> </user_list> </users> </sigs>
Which means currently you have no data. And I guess it no longer blows up on you?
Hi Daniel, The SELinux Troubleshot crash does not occur all the time. It happens in a specific situation, but I could identify exactly the steps to explain to you. Today it crashed again. Let me try to explain what I did: - In the first time I see the SE Linux alert icon on the top panel, I click on it to open the SE Linux browser. - I reviewed the error and clicked ignore. I did not close the browser. - I continue to work and then I see another SE Linux alert on the top panel. I noticed that the browser did not update, so I click in the top panel icon. - Here is the point sometimes the crash happens, sometimes not. - When I click in the top panel SE Linux icon and the browser updates with "label" in the 3 lines (source process, attempted this access, on this file), then a crash happens if I click on the Delete button. I have attached a screenshot of the SE Browser window right before the crash.
Created attachment 472156 [details] SE Browser screen shot before the crash
Package: setroubleshoot-server-3.0.15-1.fc14 Architecture: i686 OS Release: Fedora release 14 (Laughlin) How to reproduce ----- See notes already documented in the bug 666622 Comment ----- See above.
*** This bug has been marked as a duplicate of bug 673147 ***