Bug 579103 - [abrt] crash in emacs-1:23.1-20.fc12: Process /usr/bin/emacs-23.1 was killed by signal 7 (SIGBUS)
Summary: [abrt] crash in emacs-1:23.1-20.fc12: Process /usr/bin/emacs-23.1 was killed ...
Keywords:
Status: CLOSED DUPLICATE of bug 549728
Alias: None
Product: Fedora
Classification: Fedora
Component: emacs
Version: 12
Hardware: x86_64
OS: Linux
low
medium
Target Milestone: ---
Assignee: Karel Klíč
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard: abrt_hash:8b5b7a142f169450f20f568e157...
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2010-04-02 16:27 UTC by James Ralston
Modified: 2013-03-03 23:00 UTC (History)
3 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2010-05-25 08:43:34 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
File: backtrace (48.44 KB, text/plain)
2010-04-02 16:27 UTC, James Ralston
no flags Details

Description James Ralston 2010-04-02 16:27:44 UTC
abrt 1.0.8 detected a crash.

architecture: x86_64
Attached file: backtrace
cmdline: emacs /home/qralston/.mulberry/Temporary_Files/External_Edit/MulberryExternalEdit.txt
component: emacs
executable: /usr/bin/emacs-23.1
kernel: 2.6.32.10-92.fc12.x86_64
package: emacs-1:23.1-20.fc12
rating: 4
reason: Process /usr/bin/emacs-23.1 was killed by signal 7 (SIGBUS)
release: Fedora release 12 (Constantine)

comment
-----
I use sawfish as my window manager. In window context, I have "Move window interactively" bound to Meta-Button1-Control-Click.

Whenever Emacs process is starting, if I grab the Emacs window and attempt to move it, Emacs sometimes crashes. It does not happen every time, and I've never been successful at reproducing the problem at will; it only happens when I'm not trying to reproduce it.

Mostly, Emacs tends to crash in this way when I launch it as an external editor from my mail client. However, the reason for that might be that when I launch Emacs by clicking its icon on the GNOME panel (which is on the bottom of my screen), Emacs finishes initializing before I can move the cursor back to the top of the screen and grab the window; in contrast, when I launch it from my mail client, the pointer is close enough to where the Emacs window appears that I can grab the window and move it while Emacs is still initializing.

Emacs never crashes if I wait until it has fully initialized before I attempt to move the window.

How to reproduce
-----
1. Launch Emacs.
2. Wait for the Emacs window to appear.
3. Before Emacs finishes initializing, grab and move the Emacs window.

Comment 1 James Ralston 2010-04-02 16:27:46 UTC
Created attachment 404208 [details]
File: backtrace

Comment 2 Karel Klíč 2010-05-25 08:43:34 UTC

*** This bug has been marked as a duplicate of bug 549728 ***

Comment 3 Karel Klíč 2010-05-25 08:43:34 UTC
This bug appears to have been filled using a buggy version of ABRT, because
it contains a backtrace which is a duplicate of backtrace from bug #549728.

Sorry for the inconvenience.

Comment 4 James Ralston 2010-05-25 20:38:21 UTC
In response to comment 3: Karel, are you sure that this bug is a duplicate of bug 549728?

This crash was caused due to SIGBUS; in bug 549728, the crash occurred due to SIGSEGV. Additionally, the stack backtrace in bug 549728 is incomplete (and thus useless), while the stack backtrace I attached is complete...

Comment 5 Karel Klíč 2010-05-26 09:05:33 UTC
Hi James, 

I think it is the same bug. Getenv call crashes on Emacs startup due to memory corruption (caused by a race condition?). Unfortunately I do not see where the actual flaw is.

#3  getenv (name=0x3a3dd3b4aa "NGUAGE") at getenv.c:84
        ep_start = Cannot access memory at address 0x51f00000020037f

Thank you for the bug report, both the backtrace and comment are useful.


Note You need to log in before you can comment on or make changes to this bug.