[Rivet] [Fastjet] Problem with JADE plugin?

[Frank Siegert]

Hi Andy,

Thanks for finding this. I still wonder about the implications also for 
other analyses (e.g. all the ones with Durham jet resolutions). I guess 
the *_max versions are the ones that should be used, aren't they? That 
affects

  ALEPH_1996_S3196992 DELPHI_1996_S3430090 DELPHI_2003_WUD_03_11
  MC_JetAnalysis (-> all MC_*JETS analyses) OPAL_2001_S4553896

I will change it for them as well now unless there are objections.

Cheers,
Frank
  


Gavin Salam, Monday 28 June 2010:
> Hi Andy,
> 
> Glad to hear the problem went away so easily!
> 
> Cheers,
> Gavin
> 
> > Date: Sun, 27 Jun 2010 22:45:53 +0100
> > From: Andy Buckley <andy.buckley at ed.ac.uk>
> > To: Gavin Salam <salam at lpthe.jussieu.fr>
> > Cc: fastjet at projects.hepforge.org, Rivet
> > <rivet at projects.hepforge.org> Subject: Re: [Rivet] [Fastjet]
> > Problem with JADE plugin?
> > 
> > On 25/06/10 17:12, Gavin Salam wrote:
> >> Hi Andy,
> >> 
> >>>> You can also try exclusive_ymerge_max (but I don't think it'll
> >>>> make much difference -- I'm not sure which corresponds to what
> >>>> Jade did).
> >>> 
> >>> I tried that -- at least for the 2->3 inclusive plot it made no
> >>> difference to use the _max version of exclusive_ymerge. For the
> >>> other transitions I got segfaults -- I guess that when using the
> >>> ymerge_max(n) I need to explicitly make sure that n jets can
> >>> definitely be found in the event?
> >> 
> >> That sounds bizarre. We've got protection in there for such cases,
> >> in which case you should just get zero. Can you provide a simple
> >> example that triggers the problem?
> > 
> > I thought I had the most up to date FastJet, but I was a patch
> > version behind: updating from 2.4.1 to 2.4.2 fixed the ymerge_max
> > crashing problem. And using the _max versions of the functions has
> > fixed all the plot shapes: fantastic. Thanks, Gavin :)
> > 
> > Holger, what FastJet version were you using? It should now
> > definitely be 2.4.2 (or higher in future): I've added a configure
> > check to that effect.
> > 
> > Andy
> 
> _______________________________________________
> Rivet mailing list
> Rivet at projects.hepforge.org
> http://www.hepforge.org/lists/listinfo/rivet