segmentation fault on any error

Bug #37255 reported by ispmarin
6
Affects Status Importance Assigned to Milestone
gnuplot (Ubuntu)
Invalid
Medium
Unassigned

Bug Description

The gnuplot version installed with dapper drake flight 5 segs fault everytime some error occurs, like a plot 'file.dat' u 1:30 and 30 is not in the file, or wrong option in gnuplot >. It crashes without mercy everytime.

Revision history for this message
Aanjhan Ranganathan (aanjhan) wrote :

Thank you for your report. Its working fine on my system. Can you please specify any specific situation when it crashes. It would be very helpful in Confirming and moving forward in fixing it.

Changed in gnuplot:
status: Unconfirmed → Needs Info
Revision history for this message
Vassilis Pandis (pandisv) wrote :

Could you please provide a backtrace as well ? See https://wiki.ubuntu.com/DebuggingProgramCrash for instructions. Thanks.

Revision history for this message
Vassilis Pandis (pandisv) wrote :

The original reporter has gone missing, so I think I have to reject this since there is not enough information to reproduce it. If this is still an issue for anybody, please re-open this. Thanks.

Changed in gnuplot:
status: Needs Info → Rejected
Revision history for this message
ispmarin (ivan-if-deactivatedaccount) wrote :

Just got a backtrace of this bug. It was a little hard to reproduce, running inside valgrind or gdb does not show the bug, but attaching the process does the trick.

Revision history for this message
ispmarin (ivan-if-deactivatedaccount) wrote :

Got the backtrace.

Changed in gnuplot:
status: Rejected → Needs Info
Revision history for this message
ispmarin (ivan-if-deactivatedaccount) wrote : Gnuplot backtrace

This is the gnuplot backtrace with gdb. Working in a better one.

Revision history for this message
ispmarin (ivan-if-deactivatedaccount) wrote : Valgrind debug output

Again, running inside the valgrind gnuplot does not show any strange behavior (typing an erroneous command showed no effect, just the invalid command warning). But the error is still here, as the gdb output can show.

Revision history for this message
Vassilis Pandis (pandisv) wrote :

Wow, you are quick :) Could you also please provide clear instructions on how to reproduce this (pick your easiest case :-) ? Thanks again.

Revision history for this message
ispmarin (ivan-if-deactivatedaccount) wrote :

Any wrong command that i tested gave the same results. Like 'plott x', or 'set style dada lines', or anything that is not the right syntax in the gnuplot gave the segfault.

Revision history for this message
ispmarin (ivan-if-deactivatedaccount) wrote :

Sorry about that! I am running a amd64 (Athlon 64 3200) with Asus SLI Mobo (nvidia chipset), 120GB Sata and nvidia 6200 fx. Running dapper lts (updated from flight 4).

Revision history for this message
Aanjhan Ranganathan (aanjhan) wrote :

Does this bug still exist in Edgy or latest Dapper versions? If so can you please upload a sample file.dat file? I am unable to reproduce it on the latest Edgy package by typing 'plott x'.It says Invalid command perfectly and does not crash. I use gnuplot version 4.0.0-3 0

Revision history for this message
William Grant (wgrant) wrote :

Information was requested 7 months ago, as this bug is unreproducible. If anybody is able to reproduce this, please reopen the bug.

Changed in gnuplot:
status: Needs Info → Rejected
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.