Problem 716

Summary:	Crash after "----> Histogram Tree opened" in TestEm7
Product:	Examples/Extended	Reporter:	pshenich
Component:	electromagnetic/TestEM5	Assignee:	Gabriele Cosmo <Gabriele.Cosmo>
Status:	RESOLVED INVALID
Severity:	normal
Priority:	P2
Version:	7.0
Hardware:	PC
OS:	Linux

Description pshenich 2005-01-24 12:00:24 CET

Dear All,

I just migrated from GEANT4 6.2 to 7.0 and tried to compile and run
electromagnetic/TestEm7 "as is". I tried both PI_1_2_1-lite and even the
latest PI_1_2_4, but in both cases I have the following:

...........................
----> Histogram Tree opened
assertion failed: `false' (in
/build/moneta/SEAL_1_4_2/src/Foundation/PluginManager/src/PluginInfo.cpp:108)
[A]bort, [R]etry, [I]gnore, [C]ore, [S]tack trace? s
 0x40067ac6 _ZN4seal9DebugAids10stacktraceEi + 0x66
[/cern/PI/app/releases/SEAL/SEAL_1_4_2/rh73_gcc32/lib/liblcg_SealBase.so]
 0x400676a4 _ZN4seal9DebugAids6failedEPKciS2_S2_ + 0x304
[/cern/PI/app/releases/SEAL/SEAL_1_4_2/rh73_gcc32/lib/liblcg_SealBase.so]
 0x40035b66 _ZNK4seal10PluginInfo9noFactoryEv + 0x156
[/cern/PI/app/releases/SEAL/SEAL_1_4_2/rh73_gcc32/lib/liblcg_PluginManager.so]
 0x400c9907
_ZNK4seal13PluginFactoryIFPN11AIDA_Plugin15AIDA_PluginTypeEvEE6createENS_17MapUniqueIteratorISt8multimapISsPNS_22PluginFactoryImplTypesIS4_E4InfoESt4lessISsESaISt4pairIKSsSB_EEEEE
+ 0xb7 [/cern/PI/app/releases/PI/PI_1_2_4/rh73_gcc32/lib/liblcg_AIDA_Plugin.so]
 0x400c9a18
_ZNK4seal13PluginFactoryIFPN11AIDA_Plugin15AIDA_PluginTypeEvEE6createERKSs +
0x58 [/cern/PI/app/releases/PI/PI_1_2_4/rh73_gcc32/lib/liblcg_AIDA_Plugin.so]
 0x400cb37d _ZN11AIDA_Plugin17AIDA_PluginLoader4loadERKSs + 0x4d
[/cern/PI/app/releases/PI/PI_1_2_4/rh73_gcc32/lib/liblcg_AIDA_Plugin.so]
 0x4036923a _ZN2pi9aida_tree16AIDA_TreeFactory6createERKSsS3_bbS3_ + 0x10a
[/cern/PI/app/releases/PI/PI_1_2_4/rh73_gcc32/lib/liblcg_AIDA_Tree.so]
 0x0805d0d2 ? + 0x805d0d2 [TestEm7]
 0x083e49e3 ? + 0x83e49e3 [TestEm7]
 0x083e45e8 ? + 0x83e45e8 [TestEm7]
 0x083ed577 ? + 0x83ed577 [TestEm7]
 0x0859fc80 ? + 0x859fc80 [TestEm7]
 0x085ad636 ? + 0x85ad636 [TestEm7]
 0x085b61a6 ? + 0x85b61a6 [TestEm7]
 0x085ab8d1 ? + 0x85ab8d1 [TestEm7]
 0x085bae0f ? + 0x85bae0f [TestEm7]
 0x0859fc80 ? + 0x859fc80 [TestEm7]
 0x085ad636 ? + 0x85ad636 [TestEm7]
 0x0804fe29 __strtod_internal + 0x1059 [TestEm7]
 0x42015704 __libc_start_main + 0xe4 [/lib/tls/libc.so.6]
 0x0804fb11 _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_c + 0x41 [TestEm7]

The same example taken from 6.2 version works fine after commenting out
"delete af" in his RunAction.cc version. I guess that the problem appeared
in changes introduced into RunAction.cc of TestEm7 within the version 7.0.
Am I right ?

Many thanks in advance for your comments,
Regards,

Igor Pshenichnov

Comment 1 Michel.Maire 2005-01-26 09:33:59 CET

Michel,

  I have tested with Geant 4.7.0 and PI 1_2_3.
  I ran TestEm7 with proton.mac, there was no problem.
  Program was finished without error and testem7.paw was created.

This bug seems to come from PI.
I had a similar problem with PI _1_2_3, and it came from  a missing library on
my system: libshift.so

The user should use the PI Savannah portal (
https://savannah.cern.ch/projects/pi/ ) to report this bugs.
Regards,    Jean Jacquemier.

Comment 2 pshenich 2005-01-27 12:07:59 CET

Dear All,

Concerning the current problem plese follow the link where I get hints from PI
experts:
https://savannah.cern.ch/bugs/index.php?func=detailitem&item_id=6447

In summary, on my local sytem, not in CERN, `ldd liblcg_AIDA_HBookStore.so`
complained about libshift.so. In standard PI installation (which includes most
of popular CERN software) libshift.so is absent since it is a component of
licensed? CASTOR software.  Getting libshift.so in LD_LIBRAY_PATH makes things
working and HBOOK files can be produced.

Regards, Igor

Comment 3 John.Allison 2005-02-04 03:13:59 CET

Not my area of expertise (assigned to me in error?).  Gabriele, you will know to whom to assign.  John.

Comment 4 Gabriele Cosmo 2005-02-04 03:37:59 CET

Following the history of the report, the problem is not Geant4 related
and can be solved in the context of PI. The report is therefore closed.