Hi again, maybe I should a little bit more specific. Consider class A { [...] TClonesArray *fA \\-> } and A::A() { fA = new TClonesArray(...); } A::~A() { delete TClonesArray(); } When I use SetAutoDelete(false), which is the default, I don't see any difference with the use of an additional static member fsA. When reading a tree and the contructor is only called once and TClonesArray is reused in the same way as in static case. The disadvantage of the static array is clearly that I cannot use two objects of this type in parallel. >From reading the documentation 3.1c I got the impression that the default c'tor (at least in the form 'new(address) Class()') is always called for initialization before an object is read in. However, this seems not to be the case. Best regards, Volker -- Dr. Volker Hejny Tel: 02461/616853 ** Institut f. Kernphysik Fax: 02461/613930 ** ---------------------------------------------------------------- ** ** --- Forschungszentrum Juelich GmbH, D-52425 Juelich **
This archive was generated by hypermail 2b29 : Sat Jan 04 2003 - 23:50:39 MET