Re: [ROOT] Creation of chains

From: Rene Brun (Rene.Brun@cern.ch)
Date: Thu Sep 19 2002 - 12:09:34 MEST


Hi Ole

Do you use a TApplication or TRint? Are you using option "-b" to start
your application?

>Is it safe to replace the "file->Write()" with"tree->GetFile()->Write()"

Simply use tree->AutoSave. The Tree knows to which file it should go.

Rene Brun

On Thu, 19 Sep 2002, Ole Streicher wrote:

> Ciao Rene!
> 
> Rene Brun writes:
>  > > Fill: Switching to new file: Run0001.Physics.Tracker.Raw_1.root
>  > > dlopen error: /opt/root/lib/root/libGpad.so: undefined symbol: __ti10TPaveLabel
>  > > Load Error: Failed to load Dynamic link library /opt/root/lib/root/libGpad.so
>  > > *** Interpreter error recovered ***
> 
>  > I do not understand this problem! Are you linking the main executable
>  > yourself or using root.exe?
> 
> Myself. My command line to link is
> 
> c++ -g /opt/root/lib/root/libHist.so    \
>        /opt/root/lib/root/libMatrix.so  \
>        /opt/root/lib/root/libTree.so    \
>        /opt/root/lib/root/libCore.so    \
>        /opt/root/lib/root/libCint.so    \
>  -o .libs/rawreader rawreader.o         \
>     ./.libs/librawpamela.so             \
>     ../event/.libs/libpamelarun.so -ldl \
>     /usr/local/lib/liblog4cpp.so -lnsl -Wl,--rpath -Wl,/usr/local/lib
> 
>  > Could you try with the recently released 3.03/09? 
> 
> I used that version.
> 
>  > Make sure to have the right $ROOTSYS/etc/system.rootrc
> 
> I have it at /etc/root/system.rootrc (where it was installed together
> with 3.03/09), but I also tried to have an /opt/root/etc/system.rootrc
> (my $ROOTSYS is /opt/root).
> 
> There is also another problem: as far as I know, I need to do a
> file->Write() for all open files before finishing the program. This
> will ofcourse fail if the tree has changed the file - I get an error
> message "<TKey::Write>: No file open" or even a segfault now. 
> 
> Is it safe to replace the "file->Write()" with "tree->GetFile()->Write()"
> for all trees that are used to write?
> 
>  > > BTW, is it possible to make the max treesize local to the tree? 
>  > Currently, the maxtreesize is global. It could be made local to the Tree.
>  > With may be some complication in case a Tree is split over several files.
> 
> The request is neither serious nor urgent; it was just an idea.
> 
> Something different: what happens if there is more than one tree in
> the file? Am I still able to chain all the trees of the files together
> respectively?
> 
> Ole
> 



This archive was generated by hypermail 2b29 : Sat Jan 04 2003 - 23:51:10 MET