cbmreco: Fix potential memleak while collecting profiling info.

Refs #3028

requested review from @p.-a.loizeau

assigned to @p.-a.loizeau

changed the description

added BugFix Online labels

Dear @d.smith, @fweig, @f.uhlig, @v.friese, @p.-a.loizeau,

you have been identified as code owner of at least one file which was changed with this merge request.

Please check the changes and approve them or request changes.

added CodeOwners label

Given that Dirk reported at least on OOM killed job during the night, I will probably merge this after !1728 (merged). Will try in the meantime to check if I can rerun the memleak tool that we used in the issue on this branch

added 4 commits

• dc359fe5...33e0b19f - 3 commits from branch computing:master
• d92cd321 - cbmreco: Fix potential memleak while collecting profiling info.

Compare with previous version

@fweig I still see the same memleak, maybe linked to the OS?

My test method on the virgo submit node (I could also add it to the redmine issue if you want):

ssh -Y vae23.hpc.gsi.de
source /lustre/cbm/users/ploizeau/apptainer/build_cbmroot_test_memleak.sh
# That should run a first test on the full file of 20TS
cd /tmp/${USER}/cbmroot_build/bin
./cbmreco -p /tmp/ploizeau/cbmroot/parameters/online/ -i /lustre/cbm/users/ploizeau/mcbm2022/2391_first20Ts.tsa -n 1
# [...] Optional debugging ;-)
cd
rm -rf /tmp/${USER}/

Should I merge nonetheless? Probably it cannot hurt right?

resolved all threads

merged