hi, for my point 1) I attached the file sesion.tar.gz and in this file there are two sesions as *.txt. sesion0.txt includes the one in which I set the options like that Root.MemStat: 1 Root.ObjectStat: 1 Root.MemCheck: 1 (these are the only ones I changed) and the sesion1.txt is the one in which the options above are set to 0. dqm4.C is the same script for two sesions. There is a difference between the sizes of the heap area for two sesions if I do not miss something. Should not they show up the same number for same script ? for my point 2) I could not re-produce the phenomena in a smal script and it is meaningless to attach whole the list of files because I am using external libraies; as soon as possible I will reproduce and post, for now we can forget about it. note : at the end of the sesions there can be seen lots of cint /tmp errors which I post that behavior in a different mail, you can forget about them in the scope of this mail. thank you oz ROOT3.5.1,CERNRetHat7.3.1,AMD Rene Brun wrote: > Hi Ozgur, > > Ozgur COBANOGLU wrote: > > > > (root3.3.9 and root3.5.1, CERN RedHat7.3.1, Intel P3) > > > > hi rooters, > > > > I have two points: > > > > point 1) while the options in the $ROOTSYS/etc/system.rootrc are set like that > > (I do not have .rootrc) > > > > Root.MemStat: 1 > > Root.ObjectStat: 1 > > Root.MemCheck: 1 > > (these are the only ones I changed) > > > > I do gObjectTable->Print() and I see that while execution the heap size > > increases. Normally I think that I have missed deleting something in one of > > the loops. But then if I set the options above to 0 and do > > gObjectTable->Print() I see that the size of the heap stays the same and of > > course the size of the list decreases. Which one should I trust? or Whats the > > difference between them ? > > > > gObjectTable->Print shows a table with the current number of objects per class. > This function cannot report the size of dynamic structures like float* > or any structure/class not known to CINT. The table itself requires some space > but should stabilize quickly after a few calls. > You do not give enough information. > > > point 2) if I dynamicly declare an object (say TPad) in a named script, the > > functions which the named script calls can not see that object after the first > > call. The first call is always succesfull but after the first call the objects > > seem to be gone from the heap ; I cant find their addresses by gDirectory or > > gROOT. CINT complains like that : > > I cannot follow your explanation. Looks like a typical C++ scoping problem. > Again, not enough information > > Rene Brun > > > > > leftUp is not defined in current scope... > > > > What is the scope rule that I do not know ? > > > > Thanks for helps.. > > > > ozgur > > -- ---------------------------------------------------------------- Özgür Çobanoglu E-Posta: ozgur@nucleus.istanbul.edu.tr istanbul Üniversitesi Tel: +90 212 511 84 80 Nükleer Fizik Bölümü Fax: +90 212 519 08 34 Vezneciler istanbul Mobil: +90 505 512 72 90 34459 Türkiye Bina: Fizik Bölümü ---------------------------------------------------------------- Özgür Çobanoglu E-Mail: Ozgur.Cobanoglu@cern.ch CERN EP/AID Tel: +41 (22) 767 2382 CH-1211 Geneva 23 Fax: +41 (22) 767 9585 Switzerland Building: 53-R-024 ----------------------------------------------------------------
This archive was generated by hypermail 2b29 : Thu Jan 01 2004 - 17:50:10 MET