[solved] DTL_CEL/partran/cli segfault

[yngvelevinsen]

Hi,

I have a DTL lattice I tried to run using the CLI (noX11) on Linux, but I get a segfault. I manage to make a minimal example attached.

It looks like if I use QUAD instead of DTL_CEL then it does not fail. It also seem to only fail when partran is turned on. Finally it works fine if I run in GUI, and it works with older CLI versions (not sure when this issue started).

The output I get is:

Code: Select all

...
   Multiparticle calculation with PARTRAN                          
                                                                          
  Element :1  [0.00000 m]  Ngood :100000                                         
 [ 0% ]                                                                   
QVariant::load: unable to load type 64.                                   
QVariant::save: unable to save type 'QFont' (type id: 64).                
                                                                          
Critical system error:                                                    
Segmentation violation

[yngvelevinsen]

I can add that e.g. version 2.20.0.0 seems to work fine, maybe that helps narrow down what changed.

[Didier]

Dear yngve,

It's fixed.

Regards,

Didier

[yngvelevinsen]

Dear Didier,

As always, thanks for the prompt response. I checked the 2.21.0.0, but I still get segmentation violation?

Code: Select all

% ../TraceWin project.ini hide_esc
   Data Reading...

 Reference particle transports in Elements: 1 
 Reference particle transports in Elements: 2 
 Total Line : 2 Elements 
  Calculation of line envelope
   : finished 
  Number of step : 35
  Memory used   : 1 kBytes
  New Data file writting :
  /storage6/public/TestTraceWin/mini_dtl/calc/lattice.dat
  New Data file writting (with detailed DTL_CEL):
  /storage6/public/TestTraceWin/mini_dtl/calc/lattice_2.dat
  Multiparticle calculation with PARTRAN
 Element :1  [0.00000 m]  Ngood :10000
Critical system error:
Segmentation violation
% ../TraceWin -v
Current version: 2.21.0.0

[Didier]

Fast maybe but not very efficient and rigorous
In fact, there were two errors that accumulated and when I corrected the first one, I didn't even test the specific case that you had given me.
So it should be good now, I hope and I didn't change the version.

Regards,

Didier

[yngvelevinsen]

Thanks Didier, now it works again!