Dear Renate Although I am not a root expert, here may be a hint: I have often experienced the same error message, and in my case I have found that I get them when I try to fill a tree, but have a dangling pointer i.e. a pointer that points to a random address. I hope this helps Christian ---------------------------------- C.h.r.i.s.t.i.a.n S.t.r.a.t.o.w.a V.i.e.n.n.a. A.U.S.T.R.I.A e-mail: cstrato@EUnet.at Renate Bornheim wrote: > Dear Rooters, > > I work on a root interface for the existing analysis frame of the CLEO > collaboration and I have a persistent problem with a fatal in operator > delete. I can't post any code since the interface of several routines > and is totally dependent on the local structure, sorry. > It's all C++ with calls to TH1, TNtuple, TDirectory etc in the end to > produce a root format output instead of hbook. > I get " Fatal in <operator delete>: storage area overwritten " when > cleaning up the internal 'new's in my interface code. > I checked that I don't do anything like delete xxx; instead of delete [] > xxx or explicit double deletion of one of my objects. > I also put several printouts in the source code, so I know this message > doesn't come from any of the destructors in routines like : TBuffer, TH1, > TDirectory ... > > I'd like to know if anybody else experienced this error message?? > .. where? how? how did you solve it ? > ANY hint could be helpful !! > It only shows up on OSF, never appears on SUN !!!!!! > I work with ROOT 2.25/03 . > > What other internal root routines are called during the cleanup phase > ? I'd like to see the routine where this fatal is acutally produced in, > since this might give me a hint what's wrong in my code. > As far as I could see the error is produced in TStorage via the MEM_DEBUG > macro ? How can I modify the source code to get a print out telling me > from which routine the 'delete' error actually comes ?? > > Interesting enough: I can also produce this error when going through a > constructor of one of the interface routines during the stage of booking > all histograms / ntuples in my test program ...that is long before the > end and cleanup phase! > What kind of internal 'delete' in ROOT could be executed at that early > stage of the program ? > > I know it's hard to say anything without seeing the code ... but thanks a > lot for trying ! > > Renate Bornheim > > _ _ > ( )_^_( ) -----------------------------------------* > { . . } Renate Bornheim > \ * / renate@mail.lns.cornell.edu > / - - \ > / \ Cornell University > (_} o {_) Wilson Lab > | _ | > (_) (_) _________________________________________*
This archive was generated by hypermail 2b29 : Tue Jan 01 2002 - 17:50:37 MET