erosita-forum

Hi,

I have successfully installed eSASS specifying gcc-12, but I get the following problem.

python3 demoscript_DR1.py /media/finoguen/Expansion/data5/erass1/ero_archive/180/001/EXP_010/em01_001180_020_EventList_c010.fits /home/finoguen/tmp
This script will use eSASS installed in None
and the executables in:
/home/finoguen/sci/eSASS4DR1/eSASS/bin.
evtool: 2.29.2.3 eSASS4DR1 Mar 14 12:47:18 2025

Program received signal SIGSEGV: Segmentation fault - invalid memory reference.

Backtrace for this error:
#0 0x630df3194bd1 in ???
#1 0x630df31941a5 in ???
#2 0x989b2e4532f in ???
at ./signal/../sysdeps/unix/sysv/linux/x86_64/libc_sigaction.c:0
#3 0x7ffe5433c4e0 in ???
Traceback (most recent call last):
File "/home/finoguen/sci/eSASS4DR1/eSASS/scripts/demoscript_DR1.py", line 536, in <module>
main(sys.argv)
File "/home/finoguen/sci/eSASS4DR1/eSASS/scripts/demoscript_DR1.py", line 528, in main
RunTasks()
File "/home/finoguen/sci/eSASS4DR1/eSASS/scripts/demoscript_DR1.py", line 122, in RunTasks
subprocess.check_call(cmd)
File "/usr/lib/python3.12/subprocess.py", line 413, in check_call
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['evtool', 'eventfiles=/media/finoguen/Expansion/data5/erass1/ero_archive/180/001/EXP_010/em01_001180_020_EventList_c010.fits', 'outfile=/home/finoguen/tmp/EventList_021.fits.gz', 'emin=0.200000', 'emax=0.600000', 'image=yes', 'rebin=80', 'size=3240', 'pattern=15']' died with <Signals.SIGSEGV: 11>.

Thanks,
Alexis

Hi Alexis, I remember we had a memory leak issue a while back but off the top of my head I cannot recall whether this was patched in all releases. We may not have been able to do so in the DR1 code. I'll check and get back to you.

Of course ideally we would like the demo script to run without error but its failure should not keep you from trying evtool on 'real' data.

Cheers Ian.

Hi Ian,
I still encounter a problem when I run it myself:
evtool
evtool: 2.29.2.3 eSASS4DR1 Mar 14 14:13:21 2025
Event files - in [] : eb01_117048_020_EventList_c010.fits
Event file - out [] : tmp.fits

Program received signal SIGSEGV: Segmentation fault - invalid memory reference.

Backtrace for this error:
#0 0x57dff785ebd1 in ???
#1 0x57dff785e1a5 in ???
#2 0x127bfc44532f in ???
at ./signal/../sysdeps/unix/sysv/linux/x86_64/libc_sigaction.c:0
#3 0x7ffced9c17c0 in ???
Segmentation fault (core dumped)

Cheers,
Alexis

Hi again Alexis,

Well I have been able to compile a patch for this issue on my laptop, but due to manpower issues, staff departures and absences and so forth, it may be a little while before we can issue a patched release, alas. There's a bunch of automake stuff that needs to be tweaked, plus creating Docker images and I don't know what else. So in the mean time all I can do is hope that the issue does not affect you too badly.

Cheers Ian.

A possible workaround might be if you could install the release under an older OS. It is newer compilers that seem to bring this issue to light. I think we ourselves only began to experience issues after we upgraded our local OS to Ubuntu 22 (jammy).

Hi Ian,
Thanks for looking into this,
We are not allowed to use other versions of Ubuntu.
I will wait for the patch.
I was wondering whether I should recompile HEADAS using the same version of gcc.
Cheers,
Alexis

Hi Alexis,

No, it will not help if you recompile HEADAS with the same version of GCC.

We are working on a patch, but it will take a few more weeks.

Cheers,
Miriam

Hi,

The patch is available now on the DR1 site.

Cheers,

Miriam