[Rock-dev] Pure virtual method called

[Chris Mueller]

Hi,

I'm afraid i have to report again some issues concerning this fix for 
the "pure virtual function segfault"

I have tested this fix with four systems and it seems it applied 
successfully for three of them.

Unfortanetely i have one problem child where every deployment seems to 
fail every second time
and i have currently no clue why.

This is the last commit for orocos.rb on that machine:

commit f7364f101f0a71f747a8a371325d835dbaddb54e
Author: Sylvain Joyeux <sylvain.joyeux at m4x.org>
Date:   Wed Oct 26 18:42:58 2011 +0200

     roby: make rock-roby BLA fail with a proper error message

The branch is selected to master.

I also already asked the owner of that machine if he also rebuilds the 
complete package
and he reconfirmed that.

Is this error still occuring on Asguard or other systems?

I think i also observed this error last week on AVALON when starting the 
supervision.
though it occured only one time within 3 hours.

Here's also the gdb output for the relating core dump from the tested 
desktop system:

Core was generated by `ruby1.8 local_system_sender.rb --freq 1 --size 1 
--connect sherpa'.
Program terminated with signal 6, Aborted.
#0  0x00abe422 in __kernel_vsyscall ()
(gdb) bt
#0  0x00abe422 in __kernel_vsyscall ()
#1  0x0013a651 in raise () from /lib/tls/i686/cmov/libc.so.6
#2  0x0013da82 in abort () from /lib/tls/i686/cmov/libc.so.6
#3  0x01ed552f in __gnu_cxx::__verbose_terminate_handler() () from 
/usr/lib/libstdc++.so.6
#4  0x01ed3465 in ?? () from /usr/lib/libstdc++.so.6
#5  0x01ed34a2 in std::terminate() () from /usr/lib/libstdc++.so.6
#6  0x01ed4155 in __cxa_pure_virtual () from /usr/lib/libstdc++.so.6
#7  0x00fb25b2 in RTT::ExecutionEngine::processChildren() () from 
/opt/software/install/lib/liborocos-rtt-gnulinux.so.2.4
#8  0x00fb3382 in RTT::ExecutionEngine::step() () from 
/opt/software/install/lib/liborocos-rtt-gnulinux.so.2.4
#9  0x01011ea1 in RTT::base::RunnableInterface::loop() () from 
/opt/software/install/lib/liborocos-rtt-gnulinux.so.2.4
#10 0x00faab98 in RTT::Activity::loop() () from 
/opt/software/install/lib/liborocos-rtt-gnulinux.so.2.4
#11 0x0100207b in RTT::os::thread_function(void*) () from 
/opt/software/install/lib/liborocos-rtt-gnulinux.so.2.4
#12 0x00a5496e in start_thread () from /lib/tls/i686/cmov/libpthread.so.0
#13 0x001dda4e in clone () from /lib/tls/i686/cmov/libc.so.6

Hope this is a little helpfull.

Chris



On 28.10.2011 18:22, Sylvain Joyeux wrote:
> On 10/28/2011 02:19 PM, Chris Müller wrote:
>> In which package / library is this fix pushed?
>>
>> I'm afraid to say on our AVALON system and on one of the Desktop PCs for
>> RIMRES, the following
>> termination is still occuring when we try to start any deployment via
>> supervision or a simple script.
>>
>> "pure virtual method called
>>
>> terminate called without an active exception"
>>
>>
>> Both system should be completely updated.
>>
>> Do you still need a core dump about that @Sylvain?
> master or next ?
>
> I just realized that I managed to revert on master the fix from next ...
> I reverted the revert and just pushed it to master.
>
> Sylvain