Re: 'How to call CERNLIB routines..': small bug and question

Rene Brun (Rene.Brun@cern.ch)
Mon, 20 Apr 1998 10:49:00 +0200

Simon Peeters wrote:
>
> Hi Rooters,
>
> When I tried to use the CERNLIB routine 'denlan', I was able to create
> some strange Landau-distributions. The problem turned out to be that the
> 'CERNfunctions.h' used:
> extern "C" Double_t denlan_(Float_t *x);
> --------
> while it should be:
> extern "C" Float_t denlan_(Float_t *x);
> -------
> since denlan from CERNLIB returns a real. I don't have a clue where/why
> this type conversion goes wrong, but it seems to do. I compared both
> 'rootCERN' versions using the following macro:

Hi Simon,
Thanks for reporting this problem. The type must be Float_t and
not Double_t.
I have modified http://root.cern.ch/HowtoCERNLIB.html

Rene Brun