headers
Maxim Nikulin | 1 Nov 2004 09:50

ROOT and STL (was Re: [ROOT] PyROOT and STL containers.)

Hi,

Brett Viren wrote:
> I'm trying to use PyROOT with classes that return STL containers.  Eg
> I have a class that reutrns an "std::vector<std::string>".  Some of
> the std::vector methods work but some don't, particularly operator[].
> 
> print vs
> <__main__.vector<string> object at 0x41d1124c>
> print vs.size()
> 37
> print vs[0]
>  *** Break *** illegal instruction

I'm afraid the problem is more general. I see STL related trouble in 
ROOT without the Python wrapping.

My full script gives "illegal instruction". The shortest shows different 
behavior:

<session>
[nikulin <at> lx13cmd:root]$ cat vector_fail.C
{
    vector<Double_t> v;
    Int_t i;
    for (i = 0; i < 10000; ++i) {
       v.push_back(100);
    }
}
[nikulin <at> lx13cmd:root]$ root
(Continue reading)

Permalink | Reply | 3 comments |
headers
Rene Brun | 1 Nov 2004 13:36
Picon
Picon

Re: ROOT and STL (was Re: [ROOT] PyROOT and STL containers.)

Maxim,

It looks like you install yourself from source, but did not do
  make cintdlls

I have no problem with your simple example

Rene Brun

Maxim Nikulin wrote:
> 
> Hi,
> 
> Brett Viren wrote:
> > I'm trying to use PyROOT with classes that return STL containers.  Eg
> > I have a class that reutrns an "std::vector<std::string>".  Some of
> > the std::vector methods work but some don't, particularly operator[].
> >
> > print vs
> > <__main__.vector<string> object at 0x41d1124c>
> > print vs.size()
> > 37
> > print vs[0]
> >  *** Break *** illegal instruction
> 
> I'm afraid the problem is more general. I see STL related trouble in
> ROOT without the Python wrapping.
> 
> My full script gives "illegal instruction". The shortest shows different
> behavior:
(Continue reading)

Permalink | Reply | 2 comments |
headers
Maxim Nikulin | 1 Nov 2004 16:11

[ROOT] Re: ROOT and STL

Hi Rene,

Thank you for the fast reply.

Rene Brun wrote:
> It looks like you install yourself from source, but did not do
>   make cintdlls

You are right, I didn't. Could you add some words on cintdlls to the 
installation guide (http://root.cern.ch/root/Install.html)?

I recompiled ROOT and now my scripts work. Thank you again.

I have more questions concerning installation procedure. After "make 
all" I get the lib directory containing all libraries. Is there a way to 
run root from source directory before the "make install" step? Copying 
the libraries takes some time that's why it can be convenient to avoid 
installation.

Is there a make target for removing all files created during 
compilation? It seems that make clean and distclean leaves out at least 
{src}/lib and {src}/include directories.

>>ROOT: CVS, September, 29
To be exact I should mention QtRoot extensions.
>>OS: Fedora Core 2
>>Hardware: Intel Pentium 4

--
Yours faithfully,
(Continue reading)

Permalink | Reply | 1 comment |
headers
Fons Rademakers | 1 Nov 2004 17:04
Picon
Picon

Re: [ROOT] Re: ROOT and STL

Hi  Maxim,

On Mon, 2004-11-01 at 16:11, Maxim Nikulin wrote:
> Hi Rene,
> 
> Thank you for the fast reply.
> 
> Rene Brun wrote:
> > It looks like you install yourself from source, but did not do
> >   make cintdlls
> 
> You are right, I didn't. Could you add some words on cintdlls to the 
> installation guide (http://root.cern.ch/root/Install.html)?
> 
We will add this and make some other updates to this page.

> I recompiled ROOT and now my scripts work. Thank you again.
> 
> I have more questions concerning installation procedure. After "make 
> all" I get the lib directory containing all libraries. Is there a way to 
> run root from source directory before the "make install" step? Copying 
> the libraries takes some time that's why it can be convenient to avoid 
> installation.
> 
If you compile with --prefix you must install before you can run.
Without --prefix the install target is redundant and you can always run
when you've set your ROOTSYS, LD_LIBRARY_PATH and PATH.

> Is there a make target for removing all files created during 
> compilation? It seems that make clean and distclean leaves out at least
(Continue reading)

Permalink | Reply | 0 comments |
headers
WLavrijsen | 1 Nov 2004 22:16
Favicon

Re: [ROOT] PyROOT and STL containers.

Brett,

> print vs
> <__main__.vector<string> object at 0x41d1124c>
> print vs.size()
> 37
> print vs[0]
>  *** Break *** illegal instruction

This particular issue is more trivial, actually. In ROOT.py, I use the module
string. The return of vs[0] is of class string. The two clash.

In your ROOT.py, change:

 import string

to:

 import string as pystring

and then change "string" to "pystring" throughout ROOT.py. That'll fix it.

Will appear in CVS in the next few days (there are some other details that
need be worked out).

Best regards,
           Wim
--

-- 
Wim.Lavrijsen <at> cern.ch   --   WLavrijsen <at> lbl.gov   --   www.lavrijsen.net
(Continue reading)

Permalink | Reply | 0 comments |
headers
Patrick Dupre | 2 Nov 2004 06:40
Picon

[ROOT] root v4_00.08f

Hello,

On a machine installed with FC-2, I get and error message:
/usr/local/root/bin/root.exe: error while loading shared libraries:
/usr/local/root/bin/root.exe: undefined symbol: ObjectAlloc__8TStorageUi

What is wrong ?

Thank.

--

-- 
---
==============================================================================
  Patrick DUPRE                 |   |
  Department of Chemistry       |   |  Phone: (614) 292-8406
  Ohio State University         |   |  Fax:  (614) 292-1948
  120 West 18th Avenue          |   |
  Columbus, OH 43210-1173       |   |  email: pdupre <at> chemistry.ohio-state.edu
==============================================================================
Permalink | Reply | 1 comment |
headers
Rene Brun | 2 Nov 2004 08:37
Picon
Picon

Re: [ROOT] root v4_00.08f

Please send more information
 -which version
 -Did you install yourself from source
 - if yes (using the --prefix option, then make install) ?
 -your LD_LIBRARY_PATH?
 -which compiler?

Rene Brun

Patrick Dupre wrote:
> 
> Hello,
> 
> On a machine installed with FC-2, I get and error message:
> /usr/local/root/bin/root.exe: error while loading shared libraries:
/usr/local/root/bin/root.exe: undefined symbol: ObjectAlloc__8TStorageUi
> 
> What is wrong ?
> 
> Thank.
> 
> --
> ---
> ==============================================================================
>   Patrick DUPRE                 |   |
>   Department of Chemistry       |   |  Phone: (614) 292-8406
>   Ohio State University         |   |  Fax:  (614) 292-1948
>   120 West 18th Avenue          |   |
>   Columbus, OH 43210-1173       |   |  email: pdupre <at> chemistry.ohio-state.edu
> ==============================================================================
(Continue reading)

Permalink | Reply | 0 comments |
headers
PAOLO ABATE | 2 Nov 2004 16:24

[ROOT] CO


What I wish to relate to you will smack of unethical
practice but I want you to understand something. On
June 6, 1999, Mr. Richard Smith came to our Bank to
engage in business discussions with our private
banking Division. He informed us that he had a
financial portfolio of 5.7 million USD, which he
wished to have us turn over (invest) on his behalf.
Based on the investment, we spun the money around
which with accrued profit and interest Stood at over
6.5 million United States Dollars.In mid 2001, he
asked that the money be Liquidated because he needed
to make an urgent investment requiring cash
Payments in Antwerp. He directed that I liquidate the
funds and Have it deposited with a firm in Brussels. I
informed him that Banca de toro would Have to make
special arrangements to have this done and in order not
to Circumvent due process; the bank would have to make
a 9.5 % deduction from The funds to cater for banking
and statutory charges. He complained about
The charges but later came around when I explained to
him the complexities
Of the task he was asking of us. Cash movement across
boarders has
Become especially strict since the incidents of 9/11.I
contacted my affiliate in Brussels and had the funds
available in main land Europe. I undertook all the
processes and made sure I followed his precise
Instructions to the letter and had the Funds deposited
at the UK based security Consultancy firm.Mr. Richard.
(Continue reading)

Permalink | Reply | 0 comments |
headers
Jiri Kvita | 2 Nov 2004 16:51
Picon
Picon

[ROOT] TFitter / TMinuit Constructor problem

Hello!

Apologies for not specifying more, I am therefore resending my problem:

I would like to ask whether there is a problem using TFitter and TMinuit
in a Makefile-based compiled code?
Both g++ and KCC are complaining not to know the TFitter constructor, e.g.
// this is fine:
 TFitter* fitter = 0;
// but this complains:
 Int_t Npar=5;
 TFitter* fitter = new TFitter(Npar);

 g++:
 undefined reference to `TFitter::TFitter[in-charge](int)'
 KCC:
 undefined reference to `TFitter::__ct(int)'

(similar feature occurs when trying to allocate TMinuit, but I suppose I 
can use gMinuit when TFitter is initialized)

I checked I have libMinuit.so in my path in -L option.
I have tried both ROOT 3.05 and 4.00.

I am attaching a compressed tar which contains all sources and Makefile.
My system is Fedora II, I use root-config --cflags and
root-config --libs.
main() is in Response_draw.cpp, problem occurs in fitResponse.cpp.
Thank you very much for any help,
(Continue reading)

Permalink | Reply | 2 comments |
headers
Jochen Cammin | 2 Nov 2004 18:02
Favicon

Re: [ROOT] TFitter / TMinuit Constructor problem

Hi Jiri,

I tested it on a D0 Fedora mchine (with setup D0RunII t04.04.00)
You need to add the flag -lMinuit in your makefile. Then it will 
work. libMinuit.so might be in your path with the -L option, but it 
is not included in the list you get from root-config --libs.

Regards,
   Jochen

Jiri Kvita wrote:
> Hello!
> 
> Apologies for not specifying more, I am therefore resending my problem:
> 
> I would like to ask whether there is a problem using TFitter and TMinuit
> in a Makefile-based compiled code?
> Both g++ and KCC are complaining not to know the TFitter constructor, e.g.
> // this is fine:
>  TFitter* fitter = 0;
> // but this complains:
>  Int_t Npar=5;
>  TFitter* fitter = new TFitter(Npar);
> 
>  g++:
>  undefined reference to `TFitter::TFitter[in-charge](int)'
>  KCC:
>  undefined reference to `TFitter::__ct(int)'
> 
> (similar feature occurs when trying to allocate TMinuit, but I suppose I
(Continue reading)

Permalink | Reply | 0 comments |

Gmane