Hi, With Bjorn's statement that it could be done, and Peter's earlier identification of the const problem, I tried a little harder to get things working with ROOT 3.00.06. And succeeded! However, I think I found a possible programming bug in BrTPCClusterFinder.cxx that I'd like to ask the experts to look at: As far as I can tell, fClusterIdTable is not initialized unless it has been previously used, and then only the previously used elements are reset. I was able to trace my segfault situation to an attempt at accessing one of the uninitialized elements of this array. I know from past experience that gcc on the linuxPPC platform does not automatically initialize memory...this may be different on Intel boxes. In any case, by explicitly setting the elements of this array to -1 in the Init() member, everything seems to work fine. ...steve > > Hi Steve. I don't have much to add on the cluster side, other than to say > that I have succesfully used BRAT 1-15-1 with ROOT 3.00.06 for full > reconstruction including clustering and vertexing with TPM1. It should not > be a compatibility problem... > > About the veretx, I can say that the code you sent looks very good. The > important thing is not to forget the cuts that Hiro sent you - otherwise > the vertex just isn't good. I personally use > a. z_error <0.8 > b. z_sigma <2 > when I have good statistics, but this does not make a major difference. > > Good luck :-)
This archive was generated by hypermail 2b29 : Sat Mar 31 2001 - 13:06:56 EST