Hi Anton,
I have implemented your suggestion as a new TF1 constructor:
TF1::TF1(const char *name, Float_t xmin, Float_t xmax, Int_t npar)
where name is the name of a preloaded CINT function.
In this particular case, one single name is sufficient (the TF1 name is
the CINT function name)
Rene Brun
Anton Fokin wrote:
>
> Hi Rene,
>
> that's exactly how I handle this now, but whay don't you want to have TF1
> constructor which takes const char* FunctionName as the second argument and
> assigns fMethodCall in TF1 according to this preloaded function? I think
> many user may would like to work with prloaded user defined functions in
> compiled code and be able to wrap TF1 nice facility on top of them.
>
> Best,
> Anton
>
> -----Original Message-----
> From: brun@pcbrun.cern.ch [mailto:brun@pcbrun.cern.ch]On Behalf Of Rene
> Brun
> Sent: Monday, February 28, 2000 4:34 PM
> To: Anton Fokin
> Cc: roottalk@pcroot.cern.ch
> Subject: Re: TF1 with CINT preloaded function
>
> Hi Anton,
>
> I suggest to create a compiled function like
> Double_t(*func)(double *, double *) {
> Double_t result;
> fMethodCall->Execute(result);
> return result;
> }
>
> where fMethodCall points to the preloaded CINT function.
>
> Rene Brun
>
> Anton Fokin wrote:
> >
> > Hi Rene,
> >
> > I am a bit confused. In the stress example it says
> >
> > #ifndef __CINT__
> > func = f1int;
> > TF1 *f1int = new TF1("f1int",func,-10,10,9); // this one will be
> > compiled
> > #else
> > TF1 *f1int = new TF1("f1int",f1int,-10,10,9); // this one will be
> > interpreted
> > #endif
> >
> > Of course it will work in the compiled code because #ifndef __CINT__ it
> will
> > produce correct TF1 constructor call with real pointer to a function func.
> > To get this pointer you have
> >
> > func = f1int
> >
> > with Double_t(*func)(double *, double *);
> >
> > In case of CINT it will also produce correct code because CINT will
> > recognise f1int as a name of predefined function and use it as a pointer
> to
> > a function.
> >
> > What I am asking is how to construct TF1 in the compiled code with a
> > function from preloaded CINT macro? I need to have Double_t(*func)(double
> *,
> > double *) to produce correct constructor call and I can not use the name
> of
> > a function from preloaded macro just because I am in the program, not in
> > CINT!
> >
> > Best,
> > Anton
> >
> > -----Original Message-----
> > From: brun@pcbrun.cern.ch [mailto:brun@pcbrun.cern.ch]On Behalf Of Rene
> > Brun
> > Sent: Monday, February 28, 2000 3:50 PM
> > To: Anton Fokin
> > Cc: Rene Brun; roottalk@pcroot.cern.ch
> > Subject: Re: TF1 with CINT preloaded function
> >
> > Hi Anton,
> > Look at the small example stress1 in $ROOTSYS/test/stress.cxx
> >
> > Rene Brun
> >
> > Anton Fokin wrote:
> > >
> > > Hi Rene,
> > >
> > > One of TF1 constructors is to create TF1 from a user defined function
> > >
> > > TF1 TF1(const char* name, void* fcn, Float_t xmin, Float_t xmax, Int_t
> > npar)
> > >
> > > It works fine in case of CINT
> > >
> > > Double_t myfunction(Double_t *x, Double_t *par)
> > > {
> > > Float_t xx =x[0];
> > > Double_t f = TMath::Abs(par[0]*sin(par[1]*xx)/xx);
> > > return f;
> > > }
> > > void myfunc()
> > > {
> > > TF1 *f1 = new TF1("myfunc",myfunction,0,10,2);
> > > f1->SetParameters(2,1);
> > > f1->SetParNames("constant","coefficient");
> > > f1->Draw();
> > > }
> > >
> > > Now if I want to create TF1 in my compiled code with myfunction defined
> in
> > a
> > > macro loaded via gROOT->LoadMacro() ... where can I get a pointer (void*
> > > fcn) to myfunction for TF1 constructor?
> > >
> > > Best regards,
> > > Anton
This archive was generated by hypermail 2b29 : Tue Jan 02 2001 - 11:50:20 MET