Hello!
While messing with TTrees, I've run into the following error:
Error in <TBranchElement::GetDataMemberOffset>: obsolete call with
(PhotonsList,Clusters)
The object I'm going to write in TTree is something which recall this
scheme:
// start of code
class PhotonsList: public TObject {
public:
TClonesArray* pPhotons; //photons data
PhotonsList() { pPhotons = NULL; }
~PhotonsList() {
if (!pPhotons) return;
pPhotons->Delete();
delete pPhotons;
}
void Init() { pPhotons = new TClonesArray(TObject::Class(), 8); }
ClassDef(PhotonsList,1) //list of photons for an event
}; // PhotonsList
class STreeEvent: public TObject {
public:
PhotonsList Clusters; //original photons data
void Init() { Clusters.Init(); }
ClassDef(STreeEvent,1) //event encapsulated in a tree
}; // STreeEvent
// end of code
Linking header file is:
// start of code
#ifdef __CINT__
#pragma link off all globals;
#pragma link off all classes;
#pragma link off all functions;
#pragma link C++ class PhotonsList;
#pragma link C++ class STreeEvent;
#endif
// end of code
while the object file (namely, STreeEvent.cpp) is the couple of calls to
macro to implement classes (ClassImp).
That is to say, I've a class fated to be main tree branch (STreeEvent)
which directly holds a TObject derived class, which has a pointer to a
clones-array of photons (a TObject or whatever); this pointer is not
initted in constructor since some tutorials teach so.
Now with CINT and ROOT 3.05/06 the sequence:
// loads object code to CINT:
[0] .x STreeEvent.so
// create and init an object:
[1] STreeEvent* pEvent = new STreeEvent; pEvent->Init();
// create a tree and a branch with that object as address:
[2] TTree tree; tree.Branch("Event", "STreeEvent", &pEvent);
that message prints out. It's very reasonable that a call to
TBranchElement::GetDataMemberOffset(PhotonsList,Clusters) fails, since
there is no "Clusters" field in PhotonsList and Clusters is a field /of
type/ PhotonsList (in another object).
I could bet that with 3.05/05 this did not happen, but I'm not sure.
I've no idea whether this is a message that can be safely ignored or
not.
Written enough. Any suggestion?
--
Hamlet
This archive was generated by hypermail 2b29 : Thu Jan 01 2004 - 17:50:13 MET