Problem calling DBeamHelicity_factory_CORRECTED

[sdobbs]

I tried implementing some new features, and am getting a crash that I don't understand and haven't been able to figure out what is causing it.

Basically, in DEventWriterROOT::Fill_DataTree(), I am trying to extract a DBeamHelicity object. If I get the normal, untagged DBeamHelicity object, it runs correctly. If I get the DBeamHelicity tagged as "CORRECTED", this throws the error reproduced below.

You can reproduce the error using the halld_recon branch sdobbs_beamhel_july2025_bug and the command hd_root -PPLUGINS=ReactionFilter -PReaction1=1_14__8_9_14 /work/halld/home/sdobbs/REST_add_polarization/crash/dana_rest_120777_010.hddm

12:27:05.079 [fatal] Exception in worker thread 0 while executing PhysicsEventMap2 on event #15634972
----------------------------------------------------------
JException
  Type:     unsigned long
  Message:  Unknown exception
  Function: JEventProcessor::Process
  Class:    DEventProcessor_ReactionFilter
  Plugin:   ReactionFilter
  Backtrace:

   0: void jana::components::JComponent::CallWithJExceptionWrapper const&)::{lambda()#3}>(std::__cxx11::basic_string, std::allocator >, JEventProcessor::DoLegacyProcess(std::shared_ptr const&)::{lambda()#3})
      hd_root:3d4c1f

   1: JEventProcessor::DoLegacyProcess(std::shared_ptr const&)
      hd_root:3d50fe

   2: JEventMapArrow::fire(JEvent*, std::array, 2ul>&, unsigned long&, JArrow::FireResult&)
      /u/group/halld/Software/builds/Linux_Alma9-x86_64-gcc11.5.0/jana/jana_2.4.0/Linux_Alma9-x86_64-gcc11.5.0/lib/libJANA.so:2134b0

   3: JExecutionEngine::RunWorker(JExecutionEngine::Worker)
      /u/group/halld/Software/builds/Linux_Alma9-x86_64-gcc11.5.0/jana/jana_2.4.0/Linux_Alma9-x86_64-gcc11.5.0/lib/libJANA.so:207ea9

   4: void std::__invoke_impl(std::__invoke_memfun_deref, void (JExecutionEngine::*&&)(JExecutionEngine::Worker), JExecutionEngine*&&, JExecutionEngine::Worker&&)
      /u/group/halld/Software/builds/Linux_Alma9-x86_64-gcc11.5.0/jana/jana_2.4.0/Linux_Alma9-x86_64-gcc11.5.0/lib/libJANA.so:20f20b

   5: std::__invoke_result::type std::__invoke(void (JExecutionEngine::*&&)(JExecutionEngine::Worker), JExecutionEngine*&&, JExecutionEngine::Worker&&)
      /u/group/halld/Software/builds/Linux_Alma9-x86_64-gcc11.5.0/jana/jana_2.4.0/Linux_Alma9-x86_64-gcc11.5.0/lib/libJANA.so:20f12c

   6: void std::thread::_Invoker >::_M_invoke<0ul, 1ul, 2ul>(std::_Index_tuple<0ul, 1ul, 2ul>)
      /u/group/halld/Software/builds/Linux_Alma9-x86_64-gcc11.5.0/jana/jana_2.4.0/Linux_Alma9-x86_64-gcc11.5.0/lib/libJANA.so:20f05f

   7: std::thread::_Invoker >::operator()()
      /u/group/halld/Software/builds/Linux_Alma9-x86_64-gcc11.5.0/jana/jana_2.4.0/Linux_Alma9-x86_64-gcc11.5.0/lib/libJANA.so:20effc

   8: std::thread::_State_impl > >::_M_run()
      /u/group/halld/Software/builds/Linux_Alma9-x86_64-gcc11.5.0/jana/jana_2.4.0/Linux_Alma9-x86_64-gcc11.5.0/lib/libJANA.so:20efe0

   9: ???
      /lib64/libstdc++.so.6:dbad4

  10: ???
      /lib64/libc.so.6:8a19a

  11: ???
      /lib64/libc.so.6:10f210



Closed ROOT file
Exit code: 1

[RaiqaRasool]

Hi Sean,

I looked into the issue and identified the core reason behind the crash.

As mentioned above that crash was happening on this line:

locEvent->GetSingle(locBeamHelicity, "CORRECTED");

So what is mainly happening is that above line triggers DEventSourceREST::GetObjects, which then calls DEventSourceREST::Extract_DBeamHelicity because the dataClassName == "DBeamHelicity" for this factory.

Inside Extract_DBeamHelicity, it only populates the data array with objects that satisfy getJTag() == tag. In this case, the tag for the factory is "CORRECTED", but none of the objects meet that condition. As a result, data remains empty.

However, the function still returns true even when data.size() == 0, which causes JANA to assume that the data is available and skip running the factory process. Later, when JEvent::GetSingle<DBeamHelicity> checks that data.size() != 1, it throws an exception because the size is actually zero.

To confirm this was the issue, I modified Extract_DBeamHelicity to return false when no matching data is found:

if (data.size() == 0)
    return false; // OBJECT_NOT_AVAILABLE (iter->getJTag() != tag)
else {
    factory->Set(data);
    return true; // NOERROR
}

After this change, the crash no longer occurs, since JANA now correctly runs the factory ::Process when no matching objects are found via GetObjects.