Hi Christan, Thanks for the suggestions. I should have picked up on Erik's message. By the way, your suggestion of getting some sort of search mechanism for the archive is excellent. Although Erik's message is recent enough AND his subject line is sufficiently explicity that I should have remembered seeing if (unfortunately, before I started along the same path...), I am finding it increasing difficult to locate messages from months back that I vaguely remember being relevant to a current project, but where the point that I'm interested in is buried in a message with a non-related subject. Regards, Steve Christian Holm Christensen wrote: > Hi Steve et al, > > On Mon, 30 Jul 2001 21:39:05 -0500 > "Stephen J. Sanders" <ssanders@ku.edu> wrote > >> Because the programs that I had used previously were not working >> with brat2, I started issuing some of the basic commands >> interactively to see how far I could get (I've edited out lines >> where I made typing mistakes...): > > > Good idea. Another trick is to use GDB. I somethimes compile ROOT > with debugging enabled (configure option --build=debug), so that I can > step into ROOT stuff and trace segmentation violations and so on. I > don't recommend compiling ROOT with debug symbols for production > though. > >> brat [0] BrGeantInput * gInput = new BrGeantInput("GeantInput") > > > Why don't you grab from the history file rather than screen, so that > the full lines are viasible, and do try to declare your variables, if > for nothing else, then for the ease of reading - thanks. > > >> The problem is that this is the ONLY event found (the brag run was >> for 10000 events, Au+Au at 200AGeV with 10fm<=b<=20fm), and the >> number of hits and tracks look much more consistent with 10000 >> peripheral events than 1 event at 11.12fm. It also took several >> minutes for the computer to analyze this one event, during which >> time my system "froze"--a situation I usually encounter when the >> physical memory is exhausted and I start using the swap area. > > > Erik have seen a similar problem, see: > > http://www.rhic.bnl.gov/brahms/WWW/private/list_hyper/brahms-dev-l/1141.html > > which no one seemed to take seriously. > > I did a similar test: > > BrGeantInput* io = new BrGeantInput("io"); > io->SetIOMode(BrIOModule::kBrJobFile|BrIOModule::kBrReadFile); > io->AddFile("HijingTest50ev.cdat"); > io->Init(); > io->Begin(); > BrEvent* ev = new BrEvent("ev", 0, 0); > io->Event(ev); > ev->Print("r"); // very long list > delete ev; > ev = new BrEvent("ev", 0, 0); > io->Event(ev); > ev->Print("r"); // Not so long list > io->IsEof(); // returns 1 :-( > io->IsError(); // returns 0 > > >> This looks like a possible maping problem between brag and brat. > > > I don't think so. Rather, it seems like a problem in BrGeantInput, > but I've no idea were. You see, it's simply because BrGeantInput > thinks it's at the end of the file, when in fact it isn't. I know > that there was a bit of tampering with the gbr2c.f recently, and I > don't know if those changes has propagated to BrGeantInput. > >> I will try to track it down tomorrow, but since I am doing pretty >> basic stuff I wanted to check first if someone else has actually >> seen this work with brat2. > > > Good luck. If I were you, I'd start looking for possible > uninitialised member and so on, mistakes in the Event method, and so > on. > > Yours, > > Christian ----------------------------------------------------------- > Holm Christensen Phone: (+45) 35 35 96 91 > Sankt Hansgade 23, 1. th. Office: (+45) 353 25 305 > DK-2200 Copenhagen N Web: www.nbi.dk/~cholm > Denmark Email: cholm@nbi.dk
This archive was generated by hypermail 2b30 : Tue Jul 31 2001 - 10:17:53 EDT