egs_chamber crashes when reading in cross-section data

[MartinMartinov]

After pulling recent changes (February's commits) and recompiling my HEN_HOUSE and EGS_HOME areas, I'm finding my egs_chamber simulations crash when reading certain data files, namely spinms.data and the rayleigh data set (egslog doesn't name which file, but it happens when rayleigh scattering is turned on). It happens with many of the input files that have worked over the last year. I also tried running example1.egsinp on cavity and got the same error. This is the pre-crash output:

Reading spin data base from /home/martinov/HEN_HOUSE_github/EGSnrc/HEN_HOUSE/data/spinms.data
EGSnrc spin data, version 2.0
Data generated on a machine with 1234 endianess
The endianess of this CPU is 1234
Ranges:      1.00   100.00  0.30054  1.00000


  medium    1 ..................... Segmentation fault

I checked my previous log files from before the reinstall, and I find that the output should look like this:

Reading spin data base from /home/martinov/HEN_HOUSE_github/EGSnrc/HEN_HOUSE/data/spinms.data
EGSnrc spin data, version 2.0
Data generated on a machine with 1234 endianess
The endianess of this CPU is 1234
Ranges:      1.00   100.00  0.30054  1.00000


  medium    1 .....................  done
  medium    2 .....................  done
  medium    3 .....................  done
  medium    4 .....................  done
  medium    5 .....................  done

Any ideas on what my issue is?

Config flags that might be important:

F77 = gfortran
F77_LINK = gfortran
FCFLAGS = -fPIC
FOPT = -O2
FDEBUG = -g
FLIBS =
FOUT = -o
FOBJE = o

CC = gcc
C_FLAGS = -O2 -fPIC
COUT = -o

CXX = g++
opt = -O3 -ffast-math
DEF1 =  -fPIC

Compiler version:

gcc version 4.4.7 20120313 (Red Hat 4.4.7-4) (GCC)

[rtownson]

This was introduced by the rest mass initialization fix yesterday. PR #261 is now pending.

[ftessier]

This is because we omitted to update the common block definition in the C interface as well to initialize the rest mass. Working on it:

egs_c_interface1.macros
egs_c_interface2.macros

[ftessier]

Checking with grep -iR "replace.*useful/" HEN_HOUSE/* indicates there are no other independent declaration of the COMIN/USEFUL macro in the repository:

HEN_HOUSE/interface/egs_c_interface1.macros:REPLACE {;COMIN/USEFUL/;} WITH {
HEN_HOUSE/interface/egs_c_interface2.macros:REPLACE {;COMIN/USEFUL/;} WITH {
HEN_HOUSE/src/egsnrc.macros:REPLACE {;COMIN/USEFUL/;} WITH

[ftessier]

This was fixed in pull request #261 and cherry-picked to master via #262.

🥇 to @MartinMartinov for reporting the issue.

[MartinMartinov]

Well done! Really good turn around time for this fix, I'll try it first thing in the morning.