Message: Re: Problem running example A01app Not Logged In (login)
 Next-in-Thread Next-in-Thread
 Next-in-Forum Next-in-Forum

None Re: Problem running example A01app 

Keywords: Example A01, Openscientist, histogram, OnX,
Forum: Analysis
Re: Question Problem running example A01app (Sylvia Studeny)
Date: 30 Aug, 2005
From: Guy Barrand <Guy Barrand>

   Hello Sylvia 

  The horrible author of the horrible program is back from his (long) holidays.

 Then let us try to disantangle things (if you did not throw yet
 you computer on the wall). In the below I more or less answer at once
 to your all three posts.

> Hi,
> I wanted to do some histogramm plotting and therefore tried to run 
> example A01app. I think I have installed openscientist correctly (though
> this took quite a while), at least the examples coming with openscientist
>  are running.
> When I now try to compile A01, everything goes fine (no errors).
> But I cannot run the Program:
>
>  phot:  Total cross sections from Sandia parametrisation.
>  OnX::XtUI::LookDSM_Problem :
>  Can't solve the dsm problem.
>  Bad linking order of -lXm and -lXt
>  or you try to load a DLL linked to X11 prior to
>  load a DLL linked with Motif.
>  segmentation fault
>

  This is an old sea snake problem that I had some very long time 
 ago on HP-UX. Up to now it never appeared on any Linux.

  From your further mails it seems that you attempt to install
 the binaries of the RedHat-9.0-gxx-3.2.2 or ScientificLinux-3.0.2-gxx-3.2.3
 of OpenScientist-v14r0, right ?

  As pointed out in the analysis-13, this is highly optimistic.
 To install a binary and hope that it will work at first shoot you 
 must be sure that everything is compatible. One difficult
 point with OpenScientist is that, by principle, I heavily rely 
 on local software in order to get the best performaces of the local
 machine resources. The penalty of that is around the installation. On Windows
 and MacOSX the environment does not vary too much in time and then a 
 binary installation is in general successfull. Linux is an other
 story since various "distributions" may hide various versions, or 
 worst, implementations of a given package.

  A first point is that you seem to have a g++-3.3.3 compiler. But
 even if that passes (then no compatibilty problems with the libstdc++ 
 and libc of the 3.2.3) you may expect problems coming from Motif or OpenGL.

  The upper DSM warning is clearly a symptoma that something clash around 
 Motif. The binaries had been reconstructed with an OpenMotif-2.2.3.
 You must be aware that there is another implementation of Motif, 
 called Lesstif, that may be around. Not clear to me what you have on
 your SuSE. For example it is Lesstif that is found with a cygwin-X11
 environment on a Windows.

  You may check that by jumping in the Xm.h include file that is in principle
 under /usr/X11R6/include/Xm.h. If the Copyright is from the Open Software
 Fundation ; then you have an OpenMotif else you probably have a Lesstif.

  Then before jumping to a mix with Geant4 that may add an extra
 layer of problems, I would like to be sure 
 that you have something correct for OSC(-14). I am afraid that you shall
 not be able to avoid a source installation ; it may be more painfull
 to get but at the end the situation will be much more "trustable".
 At least you must aim to have the default examples of the Lab package 
 up and running. In particular the Lab_analysis program by using 
 the "-store=xml" or "-store=root" option to produce a data file. 
 The default interactive application (by typing "onx" after having 
 the "Lab setup") must be able to read back the produced file.
 As explained in the doc, each package come with some test programs
 that permits to check a step by step installation. Note that you may
 have problems around the OnX and Lab python drivers. Not clear to me
 were python is installed under a SuSE.

  There is something that you must be fully aware here. You shall 
 deal with an analysis tool software and this kind of software 
 (whatever it is !) has, by principle, to deal with ALL computing 
 domains : IO, graphic, GUI, scripting ; and then have to embarque 
 a lot of things. Then the situation is by principle complex for this 
 kind of software. For example, Geant4 kernel
 has to deal "only" with C++, STL (and CLHEP) and you can see, by browsing
 a little bit the hypernews, that life is not so simple with "only" that.

  This being said, try to install from source by keeping your time.
 If you see that you come with nothing and are ready to throw your 
 computer on the wall (which seemed to be already done) try to get local help 
 of an experienced guy. If none is around or if you are still stucked,
 then in general at this point, I ask to the user to open me, on his machine,
 a temporary non-root account having write access on the wanted installation 
 area. (In fact now, with all these security businesses, even that may pose
 problems). You seem to be in Germany then at least the bandwidth should be ok with me.

> Something similar happens when I try example EMTest04.
>
> I'm using Geant4.7.1 on SuSe-Linux9.0 with g++ 3.3.3. Geant4 is running
> fine so far, even the A01 example (so of course, without Openscientist
> there was no histogram file).
>
> I did run the openscientist-source script and defiened G4ANALYSIS_USE
> and the CFLAGS and LIBS. I tried to run it without visualisation or
> UI-Interface, but that changed nothing.
>
> I hope youe can help me, I have tried everything I know of. Thanks a lot
> Sylvia

  Anyway, on my side I still have a Geant4.7.0 version on my machines.
 I am going to find time to install a 7.1 to be in sync with you.
 Could be find that you give me also the exact version of your CLHEP.

  I know that I had problem with A01 (I had been stucked 
 with some problem with the hadronic libs and a global kernel libs usage). 
 We can try then to concentrate on the EMTest. I never tried EMTest04. 
 I can try to rebuild it (them)....
 What is sure is that I am still with g++-3.2.3 on my Linuxes
 and don't want to change that (because 3.2.3 is what is on CERN/lxplus
 for the moment).

  About the PAW and opaw. If you want to work with PAW then 
 you need to produce an RZ files or some tuple at the ascii format.
 My OSC-14/Lab does not have an AIDA driver able to write the RZ files
 (then by using behind the old CERNLIB) or even to write some tuple ascii 
 files (I can read but not write). I shall have the writing with OSC-15 but 
 I have not yet finished the release. You may attempt to install old Analphe 
 or CERN/LCG PI to put the hand on an AIDA RZ writter (but here I can't help).
  Then here the options are :
 - try to deal with Anaphe or CERN/PI for the RZ file writing.
 - in your app try to use directly the old CERNLIB to produce the RZ files. 
   But here you shall have to deal with C++/F77 mixing. But you may already have this kind of material around.
 - try to produce, for the moment, ascii tuple files by using std::cout 
   in your application !!! PAW shall be able to read it.

  Anyway, if you finish to use PAW, I shall be interested by your kumac.
 The opaw of OSC-14 is clearly lacking some basic expected features
 of CERN/PAW but the one of OSC-15 is more compleete, in particular around the
 graphical options. I would be interested anyway by any of your ascii tuple file
 coworking by some of your .kumac in order to see if I can reemulate CERN/PAW.
 (This would be your two cents of contribution to the Geant4/analysis tools saga).

  If you intent to open me an account on your machine, the best is to 
 contact me directly : barrand@lal.in2p3.fr.

  Best regards

Inline Depth:
 1 1
 All All
Outline Depth:
 1 1
 2 2
 All All
Add message: (add)

 Add Message Add Message
to: "Re: Problem running example A01app"

 Subscribe Subscribe

This site runs SLAC HyperNews version 1.11-slac-98, derived from the original HyperNews


[ Geant 4 Home | Geant 4 HyperNews | Search | Request New Forum | Feedback ]