Re: TChain problem.

From: Axel Naumann <>
Date: Tue, 15 Feb 2005 23:53:18 -0600

Hi George,
increase the size of the TrackCalJetJCCB_trkcaljc5 array. In event 53833 it has 21 objects, which makes it overflow into the TBranch variables, which in turns makes b_AM->GetEntry fail. Cheers, Axel.

George Golovanov wrote:
> Hi Axel,
> if you can look into the: /rooms/greatlakes/ontario/gag/Calib on the clued0
> cluster there is my working dir. CalcResponse.C, Response.C and Response.h
> are the files which I compile in p14.06.01 version. CalcResponse is the
> binary. PASS2.p14.06.01.list.001 is the file list of the file which I'm
> trying to process. On the events number 53833 my program is crashed. If I
> comment
> b_EM->GetEntry(jentry);
> in the CalcResponse.C, method forEachEvent() all work successfully right up to
> last event.
> Thanks,
> - George.
> On Tue February 15 2005 11:13, you wrote:

>>Hi George,
>>in that case I'd need to look at your code. Where is it? Where's the
>>binary CalcResponse you're trying to run?
>>George Golovanov wrote:
>>>Hi Axel,
>>>	thanks for your help. What do you mean when say 12 EM object? Is it 12
>>>EM objects in event (12 values in array) or 12 leafs in branch? I have 32
>>>leafs in EM branch and each leaf usually has 3-5 values per event. I had
>>>tried to set the size of the arrays up to 100 to be on the safe side, but
>>>it doesn't help.
>>>		- George.
>>>On Mon February 14 2005 15:54, you wrote:
>>>>Hi George,
>>>>here is a guess what could cause it: You run MakeClass on a tree which
>>>>had up to 12 EM objects per branch, so MakeClass set the maximum of the
>>>>EM leaves to [12]. Now you're running on a tree which has a maximum of
>>>>14 EM objects. These don't fit into MakeClass's variables and you get
>>>>memory corruption.
>>>>Try to increase the leaf array sizes in your EM branch to say twice the
>>>>size. If that doesn't help I'd need to look at the code.
>>>>George Golovanov wrote:
>>>>>	I have the following problem. I'm trying to proccess some data in trees
>>>>>using chains. So, to constuct the program I used MakeClass and wrote
>>>>>some additional code. All root-files which I processed have the same
>>>>>structure. Basically most of the files processed good, but some of them
>>>>>rejected on any event with error like this:
>>>>>- --- *** Break *** segmentation violation
>>>>>Generating stack trace...
>>>>>0x408cf2a1 in TUnixSystem::StackTrace(void) at
>>>>>i x/src/TUnixSystem.cxx:1358 from
>>>>>i b/ 0x408ce610 in TUnixSystem::DispatchSignals(ESignals) at
>>>>>ni x/src/TUnixSystem.cxx:583 from
>>>>>i b/ 0x408cbc3a in <unknown> from
>>>>>i b/ 0x408cbcba in <unknown> from
>>>>>i b/ 0x40fd04f3 in <unknown> from
>>>>>0x4100b8b8 in <unknown> from
>>>>>0x0804a008 in <unknown> from ./CalcResponse
>>>>>0x080495cd in std::ios_base::Init::__ct( (void)) + 0x2ed from
>>>>>./CalcResponse 0x40ff9336 in __libc_start_main at
>>>>>- start.c:129 from
>>>>>/d0usr/products/rtonly/Linux-2-4/v7_1-slava/lib/ 0x080493a1 in
>>>>>s ic_ostream__tm__7_Z1ZZ2Z_RQ + 0x41 from ./CalcResponse
>>>>>Abort (core dumped)
>>>>>- ---
>>>>>This problem disappear when I comment one of the branch:
>>>>>	b_EM->GetEntry(jentry);
>>>>>in my Loop() method.
>>>>>The compilation of my code always goes successfully.
>>>>>Is this problem with data in root-files, in compilator (I use root
>>>>>v3-05-00 with KCC_4_0) or problem in my code?
>>>>>Thanks for any help,
>>>>>				    - George.

Received on Wed Feb 16 2005 - 07:01:51 MET

This archive was generated by hypermail 2.2.0 : Tue Jan 02 2007 - 14:45:05 MET