Destruction of objects when exiting root

[pliuphys]

Hi everyone. I have a class with some file.Write() in its destructor and found that if interactively I create an object and quit root, it would complain that the file is not writable. A search lead me to this post:

Warning in <TFile::Write>: file ... not opened in write mode

Hi all, on Scientific Linux SL release 5.4 (Boron), the attached macro a.C runs fine in ROOT 5.27/06b (branches/v5-27-06-patches@36515, Nov 05 2010, 15:46:56 on linuxx8664gcc) but fails to write output into the output file in ROOT 5.32/00 (branches/v5-32-00-patches@42372, Jun 14 2012, 16:18:00 on linuxx8664gcc). I run the macro in the same directory (after switching ROOT version with …/setroot.csh), so I think it is not an issue with (file) permissions, as has been the case with other posts…

a.C (520 Bytes)

I run the attachment (I added #include and using statements to make ROOT take it) in ROOT 6.10/00:

$ root

| Welcome to ROOT 6.10/00 http://root.cern.ch |
| (c) 1995-2017, The ROOT Team |
| Built for linuxx8664gcc |
| From tag v6-10-00, 13 June 2017 |

Try ‘.help’, ‘.demo’, ‘.license’, ‘.credits’, ‘.quit’/‘.q’

root [0] .L a.C+
Info in TUnixSystem::ACLiC: creating shared library /home/pliu/sandbox/dtor/./a_C.so
root [1] a obj
a ctor
f1 Writable = 1
(a &) @0x(some memory address)
root [2] .q
a dtor
f1 Writable = 0
Warning in TFile::Write: file a.root not opened in write mode
$
And there is no histogram in a.root. pcanal said “This problem will be resolved in the context of ROOT 6 (and the replacement of CINT by Cling)”. I wonder if this is actually the case?

[Danilo]

Hi,

thanks for reporting. This is a tear-down issue. I see some workaround possible, for example you could explicitly allow flushing of the content of your object with a method, perhaps also invoked in the destructor for automatic flushing when the class is not used at the prompt.

Cheers,
D