[Rivet] Rivet 1.5.0 now?

[Frank Siegert]

Hi Andy,

I actually have one last-minute issue: If a Profile1D is to be filled 
with a distribution that can be negative, there is a bug in its 
BinHeight function which doesn't allow for negative y. I suppose this 
was put in to avoid division by zero, so I propose the following patch:

Index: include/LWH/Profile1D.h
===================================================================
--- include/LWH/Profile1D.h	(revision 3000)
+++ include/LWH/Profile1D.h	(working copy)
@@ -340,7 +340,7 @@
     */
    double binHeight(int index) const {
      double bH = 0.;
-    if (sumw[index+2] > 0. && sumyw[index+2] > 0.)
+    if (sumw[index+2] != 0. && sumyw[index+2] != 0.)
        bH = sumyw[index+2]/sumw[index+2];
      return bH;
    }

(unless anybody knows why this >0 restriction should be in there, the 
code is from Lars according to SVN)

This would allow me to add a W polarisation analysis (MC_WPOL) which 
contains profile histos that have negative distributions.

Cheers,
Frank


On 09/03/11 00:07, Andy Buckley wrote:
> Hi everyone,
>
> I think that we've ironed out all the Rivet bugs, niggles, and minor
> complaints that we're aware of. Anyone else got any last minute
> problems? If not, let's release asap.
>
> I'm going to be busy all day tomorrow with interviewing, but tagging and
> tarring up this long-awaited release would make my day :)
>
> Thanks to everyone who contributed -- it always amazes me how our little
> package gets better and better with each release!
>
> Andy
>