Hi Manuel, Thanks for reporting these two problems. Both are now fixed in my development version 2.24/06. Rene Brun On Wed, 7 Jun 2000, Manuel Sanchez wrote: > Hi Rooters, > > These are two oddities (bugs?) in the code (ROOT 2.24/05): > > 1) for SetBranchStatus: > > When the status of a branch (A) in a TClonesArray is altered the > corresponding LeafCount() is activated independently of what we are doing to > the branch (A). IMHO this LeafCount() branch should only be activated when (A) > is being activated but not when it (A) is being deactivated. Otherwise > one has to write some weird code in order to deactivate all branches in a > TClonesArray, which is needed under certain circumstances. (namely, trying to > read a Tree with not all the classes in the Tree's TClonesArrays being added to > the dictionary) > > Or am I missing something? > > (P.S: Maybe it is not obvious but with "activate" I mean setting status to 1 > and deactivate means setting status to 0) > > 2) for TChain::LoadTree > > When a entry is asked which is not available in all the trees of the > chain (i.e. entryNumber > chain.GetEntries() ). The LoadTree() function deletes > the currently file opened and returns an error code. That's perfect, but when > deleting the file the "fFile" pointer is not set to 0, therefore in the > destructor of the class > > if (fFile) delete fFile; > > calls the TFile destructor a second time delivering a SEGV > > > Best Regards, > Manuel > > > --------------- > Manuel Sanchez Garcia (manuel@fpddux.usc.es) > Univ. Santiago de Comp. >
This archive was generated by hypermail 2b29 : Tue Jan 02 2001 - 11:50:27 MET