Hi Victor, Matevz, We are planning to protect the two globals gFile and gDirectory in the same way that we did we gPad. Rene Brun Victor Perevoztchikov wrote: > > Hi Matevz, > > > place to lock/unlock mutex that would prevent my threads from changing > > global variables (like gDirectory when adding new histograms). > > the problem is not so simple. The numerous globals in ROOT are very hard for > threads, even if all of them are mutex-ed > > Example: > > Thread 1 open TFile1; > Thread 2 open TFile2; > Thread 1 create histogram. > > Thread 1 thinks that its histogram in TFile1, but it is in TFile2, because > global gFile was unpredictably changed. > Mutex can not help here. > > But in the next run it could be in TFile 1 or even in TFile 9. > > So, for normal user TThread is very dangerous. It was intended to use only > for online, where users much more familiar with such problems. > > One global, gPad is defenced against it, each thread has own gPad, > theoretically it is possible to do with others, but ... > > Victor > > Matevz Tadel wrote: > > > > Hi, > > I'm running MyRint : public TRint within a thread. Where is the proper > > place to lock/unlock mutex that would prevent my threads from changing > > global variables (like gDirectory when adding new histograms). > > I tried by redefining TApplication::Start/StopIdleing(), but i'm not > > completely sure this is sufficient ... Could someone please elaborate on > > the problem ... > > > > Thank you, > > matevz > > > > --------------------------------------------------- > > Matevz Tadel, E-mail: Matevz.Tadel@ijs.si > > Department of Experimental High Energy Physics F9 > > Jozef Stefan Institute, Jamova 39, P.o.Box 3000 > > SI-1001 Ljubljana, Slovenia > > Tel.: +386-61-177-3674 Fax: +386-61-125-7074 > > --------------------------------------------------- > > -- > Victor M. Perevoztchikov perev@bnl.gov perev@vxcern.cern.ch > Brookhaven National Laboratory MS 510A PO Box 5000 Upton NY 11973-5000 > tel office : 631-344-7894; fax 631-344-4206; home 631-345-2690
This archive was generated by hypermail 2b29 : Tue Jan 02 2001 - 11:50:19 MET