VMD-L Mailing List

From: Axel Kohlmeyer (akohlmey_at_gmail.com)
Date: Mon Nov 28 2011 - 10:21:20 CST

>> it is quite possible, that you have triggered some
>> internal and hard to track down VMD bug.
>>
> I am glad I could help ;-)
>
>> i see two things that you can do:
>> a) launch VMD with -debug and then trigger
>> the "hang" again and then do a kill -STOP
>> to the "stuck" VMD process and try to get
>> a stack trace so that we get an idea what
>> is going on.
>>
> Program received signal SIGSTOP, Stopped (signal).
> 0x00007ffff633f0ae in Tcl_ParseBraces () from /usr/lib64/libtcl8.5.so
> (gdb)
> (gdb) backtrace
> #0  0x00007ffff633f0ae in Tcl_ParseBraces () from /usr/lib64/libtcl8.5.so
> #1  0x00007ffff633f822 in Tcl_ParseCommand () from /usr/lib64/libtcl8.5.so
> #2  0x00007ffff6340086 in ?? () from /usr/lib64/libtcl8.5.so
> #3  0x000000000059207d in TclTextInterp::evalFile(char const*) ()

a-ha. it looks like it is still reading/processing Tcl script from a file

[...]

> and one more:
>
> Program received signal SIGCONT, Continued.
> [Switching to Thread 0x7ffff0a66700 (LWP 839)]
> 0x00007ffff6c39f3c in pthread_cond_wait@@GLIBC_2.3.2 () from
> /lib64/libpthread.so.0
> (gdb)
> Continuing.
>
> Program received signal SIGSTOP, Stopped (signal).
> [Switching to Thread 0x7ffff7faf740 (LWP 835)]
> 0x00007ffff633f0bb in Tcl_ParseBraces () from /usr/lib64/libtcl8.5.so
> (gdb) backtrace
> #0  0x00007ffff633f0bb in Tcl_ParseBraces () from /usr/lib64/libtcl8.5.so
> #1  0x00007ffff633f822 in Tcl_ParseCommand () from /usr/lib64/libtcl8.5.so
> #2  0x00007ffff6340086 in ?? () from /usr/lib64/libtcl8.5.so
> #3  0x000000000059207d in TclTextInterp::evalFile(char const*) ()
> #4  0x0000000000592991 in ?? ()

still at the same location.

> #5  0x00007ffff62c97b6 in TclInvokeStringCommand () from
> /usr/lib64/libtcl8.5.so

[...]

>> b) use a text editor to trim off all the useless
>> script code. a saved state usually is very verbose
>> but for the most part, the script code produced
>> is identical with the default, so it would not really
>> be needed. that is particularly true for the color
>> table settings. with such a lean and mean state,
>> you may not run into any problems.
>>
> Removing the color definitions as suggested by Ajasja Ljubetič helped...

it appears that this part is what was making VMD choke
and that it was either still stuck in parsing the script
code, or that you might have triggered an overflow.

stripping the save state produced script code from
stuff that is unchanged from the default values is
the most reasonable solution here.

> Anything more I can do to help?

not at this time.

thanks,
     axel.
>> cheers,
>>    axel.
>>
> Thanks,
> Bjoern
>
> 

-- 
Dr. Axel Kohlmeyer
akohlmey_at_gmail.com  http://goo.gl/1wk0
College of Science and Technology
Temple University, Philadelphia PA, USA.