Re: [Brahms-dev-l] zdc rdo module breaks for early runs

On Tue, 22 Feb 2005, Stephen J. Sanders wrote:

> Hi:
> I am having to go back and replay some of the Yr02 AuAu raw data files
> to
> try and understand some changes that seem to have occurred in the MA.
> However, the BrZdcRdoModule fails for the early runs with:
>
> Program received signal SIGSEGV, Segmentation fault.
> 0x00be0879 in BrZdcRdoModule::Event (this=0x9a758c0,
> InputTable=0xa70de28, rdoevent=0x99b5820) at BrZdcRdoModule.cxx:563
> 563           fLTdcHist[i]->Fill(digzdc_p->GetLeftTdc(i));
>
> Is this an obvious error to one of you ZDC experts?   I assume digzdc_p
> is not being
> defined, but what would have changed going to the Yr03 runs, where the
> problem
> does no exist?  If no one has an easy fix, I'll try to devise a
> workaround.

I might see the problem - the change from year 02 to year 03 is that the
dig object changes name from ZdcDIG to ZdcDig, and this is reflected in
the code like this: (early in BrZdcRdoModule::Event)

  //----- high-gain table (old setup)

    if ((digzdc_p= (BrZdcDig*) InputTable->GetObject (BRTABLENAMES
kDigZDC)))
    {
	<set adc/tdc>
    }

    //----- low-gain table (old setup)

    if ((digzdc_p= (BrZdcDig*) InputTable->GetObject (BRTABLENAMES
kDigZDCLo)))
    {
	<set acd/tdc>
   }

    //----- current setup

    if ((digzdc_p= (BrZdcDig*) InputTable->GetObject (BRTABLENAMES
kDigZdc)))
    {
	<set adc/tdc>
    }

   // Then this is what crashes:
   for (Int_t i=0; i<3; i++) {
      fLTdcHist[i]->Fill(digzdc_p->GetLeftTdc(i));
      fRTdcHist[i]->Fill(digzdc_p->GetRightTdc(i));
      fLCfdTdcHist[i]->Fill(digzdc_p->GetLeftTdcCfd(i));
      fRCfdTdcHist[i]->Fill(digzdc_p->GetRightTdcCfd(i));
      fLAdcHist[i]->Fill(digzdc_p->GetLeftAdc(i));
      fRAdcHist[i]->Fill(digzdc_p->GetRightAdc(i));
   //    fLCfdAdcHist[i]->Fill(digzdc_p->GetLeftAdcCfd(i));
//       fRCfdAdcHist[i]->Fill(digzdc_p->GetRightAdcCfd(i));
    }

The problem is that in your case the second call sets digzdc_p correctly,
but it is then overwritten by a null pointer in the 'current setup' part
(I think). The code then segfaults when it calls digzdc_p again.

Solution: Add an 'else' to the last if-call, since we only have either
DigZDC or DigZdc present at one time:

    //----- low-gain table (old setup)

    if ((digzdc_p= (BrZdcDig*) InputTable->GetObject (BRTABLENAMES
kDigZDCLo)))
    {
        <set acd/tdc>
   }

    //----- current setup

    else if ((digzdc_p= (BrZdcDig*) InputTable->GetObject (BRTABLENAMES
kDigZdc)))
    {
        <set adc/tdc>
    }

(I haven't made the change though, so you'll have to try it :-)

regards
--Bjorn



--
Bjorn H. Samset                           Phone: 22856465/92051998
PhD student, heavy ion physics            Adr:   Schouterrassen 6
Inst. of Physics, University of Oslo             0573 Oslo
                              \|/
----------------------------> -*- <-----------------------------
                              /|\

_______________________________________________
Brahms-dev-l mailing list
Brahms-dev-l@lists.bnl.gov
http://lists.bnl.gov/mailman/listinfo/brahms-dev-l