Hi Yannick, in case you're running the scripts in interpreted mode, I'd really encourage you to try running them compiled. All it takes is making the scripts "plain C++" - no shortcuts, no CINT extentions, - and adding a few include files, which one can ifdef, such that it would be possible to interpret them as well. This extremely useful and teaching exercise saved me a lot of time: having my scripts compiled I could use symbolic debugger (gdb etc) to find logical inconsistencies in the code. I also can't exclude that you'll find some problems with your scripts during the compilation stage... -best, Pasha Patois Yannick wrote: > > Hi again. > > Sorry, this 'bug repport' is very vague... > > Some scripts I use are doing several fits on serveral histograms, and > with version 3.00.06 I had no problems. > > When I run them under 3.01.05, some fitted functions seems to 'jump' > from one histo to the other. At the end, when I display the histograms, > I found several copies of one fit function 'fitting' several histograms, > that *were fitted correctly with a function stored on the histo*, of > course the function that replace the right fits doesnt correspond to > what is expected. > > I'm sorry if I cant give more precision: all that comes out from quite > complex scripts and I didnt took the time to dig into it and find *how* > it happends. The only thing I can say is that 'magically' when I do a > Fit one one histo, another function, on another histogram 'copy' the > fitted values. Yes, functions have same name. > > If there was some obvious change made inthe way histos stores theyr > fitting functions between 3.00.06 and 3.01.05 maybe the one who wrote it > will find out why... I may come out with a little script showing this, > but probably not before a few weeks... > > Thanx, > > Yannick
This archive was generated by hypermail 2b29 : Tue Jan 01 2002 - 17:50:49 MET