Hi Art,
Yes it is still (possibly) wrong(for the details see the documentation of TTree::ChangeFile).
You should try:
TFile* pOutput=new TFile(fileName,"new");
write stuff
delete tree->GetFile();
Cheers,
Philippe.
-----Original Message-----
From: Arthur E. Snyder [mailto:snyder_at_slac.stanford.edu]
Sent: Thursday, June 12, 2008 3:44 PM
To: Rene Brun
Cc: Philippe Canal; roottalk_at_lxbuild091.cern.ch
Subject: Re: [ROOT]TFile crash in ~TFile
I was previously doing
TFile output(fileName,"new");
write stuff
then leave scope.
This is what crashed.
and changed it to
TFile* pOutput=new TFile(fileName,"new");
write stuff
delete pOutput
leave scope.
Is that wrong?
-Art S.
A.E. Snyder, Group EC \!c*p?/ SLAC Mail Stop #95 ((. .)) Box 4349 | Stanford, Ca, USA, 94309 '\|/` e-mail:snyder_at_slac.stanford.edu o phone:650-926-2701 _ http://www.slac.stanford.edu/~snyder BaBar FAX:650-926-2657 Collaboration
On Thu, 12 Jun 2008, Rene Brun wrote:
> NO !! When creating an object with new the object is allocated in the
heap.
>
> Rene Brun
>
> Arthur E. Snyder wrote:
>> "on stack" means use "new", right?
>>
>> -Art
>>
>> A.E. Snyder, Group EC \!c*p?/
>> SLAC Mail Stop #95 ((. .))
>> Box 4349 |
>> Stanford, Ca, USA, 94309 '\|/`
>> e-mail:snyder_at_slac.stanford.edu o
>> phone:650-926-2701 _
>> http://www.slac.stanford.edu/~snyder BaBar
>> FAX:650-926-2657 Collaboration
>>
>>
>>
>> On Thu, 12 Jun 2008, Philippe Canal wrote:
>>
>>> Hi,
>>>
>>> See TTree::ChangeFile which may be triggered in your case.
>>> You may have to increase increase the default limit
>>> TTree::SetMaxTreeSize(very_large_number);
>>> or allocate the TFile on the stack reather than heap.
>>>
>>> Cheers,
>>> Philippe.
>>>
>>> -----Original Message-----
>>> From: owner-roottalk_at_root.cern.ch [mailto:owner-roottalk_at_root.cern.ch]
On
>>> Behalf Of Arthur E. Snyder
>>> Sent: Thursday, June 12, 2008 10:16 AM
>>> To: Arthur E. Snyder
>>> Cc: roottalk_at_lxbuild091.cern.ch
>>> Subject: [ROOT]TFile crash in ~TFile
>>>
>>> I get a segmentation violation in ~TFile as it is trying to close my
>>> file. The trace back is below.
>>>
>>> The line pointed to in my code is 985. However, the error appears to not
>>> during |write| but when the TFile is destroyed on leave this scope at
the
>>> "}".
>>>
>>> Other identical jobs run fine. This is the largest data set processed.
It
>>> worked previously; here I've just added a few more items to output
n-tuple
>>> for a total of 32. There are 47G available where it's being written
which
>>> should be plenty and in any case it's seg violation not a out of space
>>> error.
>>>
>>> Any ideas?
>>>
>>> -Thanks, Art S.
>>>
>>> code snip:
>>> for(Int_t i=0; i<nHist; i++) {
>>> if(h[i]==0) continue;
>>> h[i]->Write();
>>> }
>>>
>>> //write out the n-tuple
>>> tup->write(); <---line 985
>>> }
>>>
>>> trace back:
>>> *** Break *** segmntation violation
>>> Using host libthread_db library "/lib/tls/libthread_db.so.1".
>>> Attaching to program: /proc/26766/exe, process 26766
>>> [Thread debugging using libthread_db enabled]
>>> [New Thread 1073746912 (LWP 26766)]
>>> 0x0394e51e in __waitpid_nocancel () from /lib/tls/libc.so.6
>>> Thread 1 (Thread 1073746912 (LWP 26766)):
>>> #0 0x0394e51e in __waitpid_nocancel () from /lib/tls/libc.so.6
>>> #1 0x038e31d4 in do_system () from /lib/tls/libc.so.6
>>> #2 0x01cd7d7f in system () from /lib/tls/libpthread.so.0
>>> #3 0x00948d6e in TUnixSystem::Exec (this=0x9ab1430,
>>> at unix/src/TUnixSystem.cxx:1768
>>> #4 0x00949208 in TUnixSystem::StackTrace (this=0x9ab1430)
>>> at unix/src/TUnixSystem.cxx:1953
>>> #5 0x00947147 in TUnixSystem::DispatchSignals (this=0x9ab1430,
>>> sig=kSigSegmentationViolation) at unix/src/TUnixSystem.cxx:936
>>> #6 0x00945369 in SigHandler (sig=kSigSegmentationViolation)
>>> at unix/src/TUnixSystem.cxx:338
>>> #7 0x0094bcda in sighandler (sig=11) at unix/src/TUnixSystem.cxx:3142
>>> #8 <signal handler called>
>>> #9 0x039db79b in main_arena () from /lib/tls/libc.so.6
>>> #10 0x007c4b5a in TDirectory::WriteKeys (this=0xbffd6ac0)
>>> at base/src/TDirectory.cxx:1951
>>> #11 0x007c30a6 in TDirectory::SaveSelf (this=0xbffd6ac0, force=false)
>>> at base/src/TDirectory.cxx:1452
>>> #12 0x007c2eda in TDirectory::Save (this=0xbffd6ac0)
>>> at base/src/TDirectory.cxx:1413
>>> #13 0x007bfa20 in TDirectory::Close (this=0xbffd6ac0)
>>> at base/src/TDirectory.cxx:500
>>> #14 0x007cd7d0 in TFile::Close (this=0xbffd6ac0, option=0xc8b3ba "")
>>> at base/src/TFile.cxx:716
>>> #15 0x007cba9c in TFile::~TFile () at include/TInetAddress.h:69
>>>
>>> #16 0x014da328 in HiggsFinder::Terminate (this=0xa546af0)
>>> at HiggsFinder.cc:985 <----mycode
>>>
>>> #17 0x05735dd3 in TTreePlayer::Process (this=0xab211d8,
>>> selector=0xa546af0,
>>> option=0x4020330 "", nentries=1234567890, firstentry=0)
>>> at treeplayer/src/TTreePlayer.cxx:2664
>>> #18 0x03f77184 in TTree::Process (this=0xae4c190, selector=0xa546af0,
>>> option=0x4020330 "", nentries=1234567890, firstentry=0)
>>> at tree/src/TTree.cxx:4793
>>> #19 0x03f4ba53 in TChain::Process (this=0xae4c190, selector=0xa546af0,
>>> option=0x4020330 "", nentries=1234567890, firstentry=0)
>>> at tree/src/TChain.cxx:1750
>>> #20 0x03fc9223 in G__G__Tree_224_0_56 ()
>>> from
>>>
/afs/slac.stanford.edu/g/babar/package/root/5.14-00e/Linux24SL3_i386_gcc323-
>>> noOptimize-Debug/lib/libTree.so
>>> #21 0x00f13ca5 in Cint::G__ExceptionWrapper (
>>> funcp=0x3fc8f74 <G__G__Tree_224_0_56(G__value*, char const*,
>>> G__param*, int)>, result7=0xbffe43e0, funcname=0xa176658 "\001",
>>> libp=0xbffe0e00, hash=0)
>>> at cint/src/Api.cxx:355
>>> #22 0x00fd9922 in G__call_cppfunc (result7=0xbffe43e0, libp=0xbffe0e00,
>>> ifunc=0xa176658, ifn=0) at cint/src/v6_newlink.cxx:465
>>> #23 0x00fc3b04 in G__interpret_func (result7=0xbffe43e0,
>>> funcname=0xbffe3fe0 "Process", libp=0xbffe0e00, hash=735,
>>> p_ifunc=0xa176658, funcmatch=1, memfunc_flag=1)
>>> at cint/src/v6_ifunc.cxx:5167
>>> #24 0x00fac1ab in G__getfunction (item=0xbffe6806 "Process(finder)",
>>> known3=0xbffe5c1c, memfunc_flag=1) at cint/src/v6_func.cxx:2600
>>> #25 0x0105d357 in G__getstructmem (store_var_type=112,
>>> varname=0xbffe5880 "U", membername=0xbffe6806 "Process(finder)",
>>> tagname=0xbffe4be0 "toff", known2=0xbffe5c1c, varglobal=0x10e1ac0,
>>> objptr=2) at cint/src/v6_var.cxx:4284
>>> #26 0x01052bf2 in G__getvariable (item=0xbffe6800
"toff->Process(finder)",
>>> known2=0xbffe5c1c, varglobal=0x10e1ac0, varlocal=0xbfff1cf0)
>>> at cint/src/v6_var.cxx:3129
>>> #27 0x00fa06ea in G__getitem (item=0xbffe6800 "toff->Process(finder)")
>>> at cint/src/v6_expr.cxx:1846
>>> #28 0x00f9ed46 in G__getexpr (expression=0xbffe7d90
>>> "toff->Process(finder)")
>>> at cint/src/v6_expr.cxx:1315
>>> #29 0x00ff7c5d in G__exec_function (
>>> statement=0xbffe7d90 "toff->Process(finder)", pc=0xbffe7d8c,
>>> piout=0xbffe7d80, plargestep=0xbffe7d70, presult=0xbffe8190)
>>> at cint/src/v6_parse.cxx:457
>>> #30 0x01001352 in G__exec_statement () at cint/src/v6_parse.cxx:4211
>>> #31 0x00ffc53f in G__exec_if () at cint/src/v6_parse.cxx:2480
>>> #32 0x01000db2 in G__exec_statement () at cint/src/v6_parse.cxx:4102
>>> #33 0x00fc5d7e in G__interpret_func (result7=0xbfff58e0,
>>> funcname=0xbfff54e0 "batch", libp=0xbfff2300, hash=514,
>>> p_ifunc=0xa415570, funcmatch=1, memfunc_flag=0)
>>> at cint/src/v6_ifunc.cxx:5954
>>> #34 0x00faca9e in G__getfunction (item=0xbfff6670 "batch()",
>>> known3=0xbfff5a8c, memfunc_flag=0) at cint/src/v6_func.cxx:2813
>>> #35 0x00fa0781 in G__getitem (item=0xbfff6670 "batch()")
>>> at cint/src/v6_expr.cxx:1859
>>> #36 0x00f9ed46 in G__getexpr (expression=0x9d7c960 "batch()")
>>> at cint/src/v6_expr.cxx:1315
>>> #37 0x00f91f67 in G__calc_internal (exprwithspace=0xbfff8f70 "batch()")
>>> at cint/src/v6_expr.cxx:203
>>> #38 0x0100878b in G__process_cmd (
>>> line=0xe56cb2 ".X
>>>
/afs/slac.stanford.edu/u/ek/snyder/trial41/analysis-42/workdir/./batch.C",
>>> prompt=0x9abdd8c "", more=0x9abdd84, err=0xbfff97cc,
>>> rslt=0xbfff97d0) at cint/src/v6_pause.cxx:2310
>>> #39 0x0089d757 in TCint::ProcessLine (this=0x9abdd68,
>>> line=0xe56cb2 ".X
>>>
/afs/slac.stanford.edu/u/ek/snyder/trial41/analysis-42/workdir/./batch.C",
>>> error=0xbfffa0e4) at meta/src/TCint.cxx:318
>>> #40 0x0089d8dc in TCint::ProcessLineSynch (this=0x9abdd68,
>>> line=0xe56cb2 ".X
>>>
/afs/slac.stanford.edu/u/ek/snyder/trial41/analysis-42/workdir/./batch.C",
>>> error=0xbfffa0e4) at meta/src/TCint.cxx:351
>>> #41 0x007a2de3 in TApplication::ProcessFile (this=0x9aff010,
>>> name=0xbfff9f23 "batch.C", error=0xbfffa0e4)
>>> at base/src/TApplication.cxx:807
>>> #42 0x007a2553 in TApplication::ProcessLine (this=0x9aff010,
>>> line=0xbfff9f20 ".x batch.C", sync=false, err=0xbfffa0e4)
>>> at base/src/TApplication.cxx:676
>>> #43 0x0062087e in TRint::Run (this=0x9aff010, retrn=false)
>>> at rint/src/TRint.cxx:328
>>> #44 0x0804894e in main (argc=1, argv=0xbfffa1c4) at
main/src/rmain.cxx:29
>>> Function batch() busy flag cleared
>>>
>>> A.E. Snyder, Group EC \!c*p?/
>>> SLAC Mail Stop #95 ((. .))
>>> Box 4349 |
>>> Stanford, Ca, USA, 94309 '\|/`
>>> e-mail:snyder_at_slac.stanford.edu o
>>> phone:650-926-2701 _
>>> http://www.slac.stanford.edu/~snyder BaBar
>>> FAX:650-926-2657 Collaboration
>>>
>>>
>>>
>>>
>>>
>>
>
>
Received on Thu Jun 12 2008 - 23:24:13 CEST
This archive was generated by hypermail 2.2.0 : Fri Jun 13 2008 - 05:50:05 CEST