Bug 4797 - Opera crashes with SIGFPE at address 087F65D3
Summary: Opera crashes with SIGFPE at address 087F65D3
Status: RESOLVED OLD
Alias: None
Product: Mageia
Classification: Unclassified
Component: RPM Packages (show other bugs)
Version: 2
Hardware: i586 Linux
Priority: Normal normal
Target Milestone: ---
Assignee: Anssi Hannula
QA Contact:
URL: https://bugs.mageia.org/show_bug.cgi?...
Whiteboard:
Keywords: UPSTREAM
Depends on:
Blocks:
 
Reported: 2012-03-05 07:26 CET by Dave Hodgins
Modified: 2012-06-10 05:12 CEST (History)
0 users

See Also:
Source RPM: opera
CVE:
Status comment:


Attachments
Backtrace from crash. (3.08 KB, text/plain)
2012-03-06 08:28 CET, Dave Hodgins
Details

Description Dave Hodgins 2012-03-05 07:26:51 CET
Going to any website that has text enterable fields
causes opera to crash when running a cauldron install.

This does not happen in a Mageia 1 install.

This has been reported to opera, with the bug id DSK-358726.

I cannot get a backtrace, as attaching gdb to the main opera
process stops opera from responding to any input.  As this
does not happen with Mageia 1, but does with cauldron, it
appears to be a cauldron problem.

While I realize opera cannot be supported by Mageia developers,
this is clearly a mageia problem.

If someone with a faster system can manage to get a backtrace,
that may help.
Comment 1 Manuel Hiebel 2012-03-06 00:28:32 CET
seems we cannot follow the upstream bug report. :s

Anssi any ideas ?

Assignee: bugsquad => anssi.hannula

Comment 2 Anssi Hannula 2012-03-06 04:17:43 CET
I cannot reproduce on 64-bit or 32-bit cauldron. You can try moving ~/.opera away or run opera as another user.

As for opera stopping responding to input, that is expected as gdb stops the target process when attaching. Use the gdb command 'cont' to resume execution.
Comment 3 Dave Hodgins 2012-03-06 08:28:09 CET
Created attachment 1685 [details]
Backtrace from crash.

Doesn't happen if I use opera -pd /home/dave/newdir

The attached backtrace is using my regular .opera directory, which
is a symlink to a directory on an encrypted filesystem on a lvm
logical valume, with around 30,000 messages accumulated over more
than a decade (originally imported to opera from a dos email pgm).

I'll forward a copy of the backtrace to opera, although it
doesn't look very useful.
Comment 4 Dave Hodgins 2012-03-27 04:25:50 CEST
Seems to be fixed in opera 11.62 Release Candidate 1, build 1347.
Comment 5 Marja Van Waes 2012-05-26 13:07:22 CEST
Hi,

This bug was filed against cauldron, but we do not have cauldron at the moment.

Please report whether this bug is still valid for Mageia 2.

Thanks :)

Cheers,
marja

Keywords: (none) => NEEDINFO

Comment 6 Dave Hodgins 2012-05-28 03:24:02 CEST
While it initially seemed to be fixed in 11.62, it just reduced the frequency.

Also occurs with opera-next, which will soon be released as opera 12.

Changing to release 2 bug, with upstream keyword.

Keywords: NEEDINFO => UPSTREAM
Version: Cauldron => 2

Comment 7 Dave Hodgins 2012-06-10 05:12:35 CEST
I finally gave in and created a new opera profile, copied the mail subdirectory,
and wand.dat, imported bookmarks and contacts, and the error no longer occurs.

No idea what was causing the error, but given I'd been using the same profile
since Mandrake 10, it had a lot of obsolete stuff.

Status: NEW => RESOLVED
Resolution: (none) => OLD


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