|
[Rivet] Rivet release: this week!?!James Monk jmonk at fnal.govMon Feb 4 18:53:55 GMT 2008
On 4 Feb 2008, at 13:17, Andy Buckley wrote: > * Class-slicing breaks projection polymorphism (say that fast 10 > times!) Fix: projection member variables -> pointers. Status? I've "fixed" it for the sole case of FastJets + VetoedFinalState (by using a pointer for the FinalState - the long-term fix will be to merge FinalState with VetoedFinalState). Are there any other cases of this? > > * A couple of projection bugs: James just reported a problem with the > jets projection compare() method, and I've been seeing a similar (?) > problem with ChargedFinalState deleting itself early. I think that turned out to be because I was using the wrong reference to a projection (one should always use the reference returned by the Event.applyProjection method). Once I started using a pointer to the FinalState in FastJets then the compare method seems to work as intended. I also saw some weird behaviour in the Particle and/or FourVector classes where data would appear to change after the momentum was accessed (like the accessor was only calculating a quantity properly on the second call or something), but I can no longer replicate it so consider it anecdotal for now. > > * HepMC version on AFS blocks super-simple use at CERN/wherever AFS is > available (I've been updating the bootstrap script to use the LCG > copies > if available). > * Composite particle veto implementation? Composite particle veto is in, although it depends on ATLAS' fix to HepMC being installed (which Lynn said would be officially released on Tuesday). The veto on a given mass combination works for me (without the HepMC patch), although I'm not sure how my combinatorics will scale up if you combine many more than two particles ;) (to be fair, I'm not sure it could ever scale). I'd like it if people who requested this feature could check it is useful in the cases they are looking at.
More information about the Rivet mailing list |