[Rivet] AGILe 1.3.0 installation bug

[Mikhail KIRSANOV]

  Hello, Andy, I fixed it.
  Sorry for this. I understand why this happened, because to obtain 
those scripts I ran the installation procedure without real reinstalling 
(in order not to check again the correctness of the installation).
  I was going to add copying the scripts to our installation Makefile, 
as well as the creation of platform links. However, I still think that the
correct place for this copying is in your build system, in the target 
install. You could just make it conditional and not do it if the 
installation path is /usr/local.
                                      Cheers, Mikhail


Andy Buckley wrote:
> (Trying again with new Genser address!)
> 
> Hi again Gensers,
> 
> We noticed during the LPCC tutorial that there is a problem with the 
> Genser installation of AGILe 1.3.0, which looks to me like the DESTDIR 
> hasn't been used during installation, so path searching is broken when 
> you relocate the binary. The easiest way to see this is via the 
> agile-config script:
> 
> abuckley at lxplus248:~$ agile-config --datadir
> /afs/.cern.ch/sw/lcg/external/MCGenerators/.work/GBUILD/external/MCGenerators_hepmc2.06.05/agile/1.3.0/share/AGILe 
> 
> 
> This should instead return something like
> 
> /afs/cern.ch/sw/lcg/external/MCGenerators_hepmc2.06.05/agile/1.3.0/share/AGILe 
> 
> 
> and the fact that it doesn't means that the use of the standard params 
> files (as specified with e.g. -P fpythia-minbias.params) doesn't work. 
> By comparison, for Rivet the correct thing appears to be being done for 
> the binary relocation:
> 
> abuckley at lxplus248:~$ rivet-config  --datadir
> /afs/cern.ch/sw/lcg/external/MCGenerators_hepmc2.06.05/rivet/1.6.0/x86_64-slc5-gcc43-opt/share/Rivet 
> 
> 
> Can the AGILe installation be fixed to use the same install scheme as 
> Rivet, please?
> 
> Thanks again,
> Andy
>