Hello all what about changing the sprintf statement: the definition is sprintf (char* dest, char* format, arg, ...); Using > Char_t crit[50]; > sprintf(&crit,"type==%d",type); means sprintf (char**, char*, ...) correct call sprintf (crit,"type==%d",type); I experienced similar behaviour: CINT seems to ignore code with incorrect pointers. My example was: >> TFile file ("fname","READ"); >> file->Close () and no error genereated by CINT; only th file wasn't closed for obvious reasons. >>>>>>>>>>>>>>>>>> Ursprüngliche Nachricht <<<<<<<<<<<<<<<<<< Am 12/6/99, 5:38:50 PM, schrieb Marco van Leeuwen <mvl@nikhef.nl> zum Thema No symbol gProofServer: > Hi, > I'm using a simple macro to plot some information from an TNtuple, read > from a file. The second time I run it, a sequence of errormessages are > generated, like "No symbol gProofServer" and "No symbol gProof". Althought > everything seems to work out well, I think this shoudl be investigated. I > tried to do it myself, but I was soon lost in the code. Therefore, I send > you my macro (see below). > Regards, > Marco van Leeuwen. > I'm using root v2.23/09 and the TNtuple in the file is not extraordinary > large (5 variables, 105 Entries). > --- Macro.C > {{ > gROOT->Reset(); > TCanvas *c1=new TCanvas("c1","Canvas #1",500,500); > gPad->DrawFrame(2.5,0.8,40,1.7); > TFile fin("vtx_cali_fit.root"); > TNtuple *NT = (TNtuple*) fin->Get("NT"); // Leaving this line out or > putting it in doesn't change anything > Char_t crit[50]; > for (Int_t > type=1; type < 4; type++) { > NT->SetMarkerStyle(type+2); > sprintf(&crit,"type==%d",type); > NT->Draw("mean:p",crit,"same"); > } > }}
This archive was generated by hypermail 2b29 : Tue Jan 04 2000 - 00:43:44 MET