[Rivet] rivet analysis path (broken --pwd option?)

[David Bjergaard]

Hi All,

Ahh, sorry for the noise, the problem was between the keyboard and the
chair.  I had forgot that $< is the first named prerequisite rather than
the last in my Makefile.  Nothing is wrong with rivet or its build
system.  

Cheers,

    Dave

David Bjergaard <david.b at phy.duke.edu> writes:

> Hi,
>
> I'm completely stymied by this one.  Hopefully some fresh/expert eyes
> can shed some light on the situation.  Here's my installation
> prescription:
>> export LOCAL_PREFIX=$HOME/rivet/local
>> mkdir -p $LOCAL_PREFIX
>> cd rivet
>> wget http://rivet.hepforge.org/hg/bootstrap/raw-file/2.1.2/rivet-bootstrap
>> chmod +x rivet-bootstrap
>> INSTALL_PREFIX=$LOCAL_PREFIX ./rivet-bootstrap
>> alias setupRivet='source ~/$LOCAL_PREFIX/rivetenv.sh; export PYTHIA8DATA=$(pythia8-config --datadir)/xmldoc'
>> mkdir -p $HOME/rivet/Analysis
>> cd $HOME/rivet/Analysis
>> # checkout analysis code from github
>> cd rivet-charmonium/
>> setupRivet 
>> make; make install
> The only thing that isn't standard rivet protocol is that I'm using a
> Makefile to build and install an external library
> (libBOOSTFastJets.so). A week ago this procedure worked great.  Now when
> I do:
>> rivet --pwd --analysis=MC_GENSTUDY_CHARMONIUM hepmc.fifo
> I get:
>> 'MC_GENSTUDY_CHARMONIUM' is not a known Rivet analysis! Do you need to set RIVET_ANALYSIS_PATH or use the --pwd switch?
>
>> There are 236 currently available analyses:
> (I've snipped the list of analyses, needless to say
> MC_GENSTUDY_CHARMONIUM isn't in the list.) 
>
> My first instinct was that there was a difference in python versions.
> I've successfully used this prescription on my laptop with python 2.7.6,
> the interactive node with python 2.7.7 built from source.  This machine
> is running 2.6.6, but I've since upgraded to 2.7.7 and still have the
> issue. I'm re-running the whole procedure on my laptop to make sure that
> this is a rivet issue and not some wonky version problem with python,
> but I'm not hopeful.  This procedure worked Monday, why is it broke now?
>
> Was there a silent change to rivet-buildplugin that makes it more
> sophisticated than calling the right g++ incantation?  
>
> Confused,
>
>     Dave
> _______________________________________________
> Rivet mailing list
> Rivet at projects.hepforge.org
> https://www.hepforge.org/lists/listinfo/rivet