[Motmot] Segmentation faults

[Andrew Straw]

Hi Jon,

The segmentation fault issue is strange... We haven't had that here.

But if you're up for a little command-line work, I think we can isolate 
it. Run fview from within gdb, the GNU debugger:

Type "gdb python" in the terminal window. After a blurb, you'll get the 
"(gdb)" prompt. Type "run /usr/bin/fview". A bunch of text will get 
printed (stuff about no debugging symbols found, etc). You'll have to 
press enter a few types at the "Type <return> to continue" text. Then, 
fview should be running as normal.

When it segfaults, however, it won't close the program but rather return 
control to gdb. At that point, type "bt" at the "(gdb)" prompt to get a 
backtrace. Send that back to this email list, especially the first 
several lines, and we'll go from there. It'll show which function is 
crashing.

-Andrew

John Schneider wrote:
> Hey Andrew,
> 
> We have run into another problem:
> Three separate computers with fresh Hardy installs, fully updated and
> the latest fview/ctrax give segmentation faults at random intervals.
> Sometimes they go for an hour or longer, other times they segfault
> within 20min. They do not seem to output errors in fview.log so I'm at
> a loss as to how to report specifics of the error.
> In terms of hardware the only common feature of all three is the e5200
> processor they have. I have run Memtest86 on them overnight and no
> errors have shown up yet so I don't think it's the RAM... Any
> input/advice in solving or diagnosing is appreciated.
> 
> Jon
> 
> PS The view->set view interval worked great (letting us use onboard
> graphics) and I checked and somehow the drive was NTFS, formatting it
> to ext3 along with 'view interval' seems to be working great.
> _______________________________________________
> Motmot mailing list
> Motmot at code.astraw.com
> http://code.astraw.com/cgi-bin/mailman/listinfo/motmot
> 


-- 
Andrew D. Straw, Ph.D.
California Institute of Technology
http://www.its.caltech.edu/~astraw/