> >// THF1 does not exist, should be TH1F
> >int typebug(THF1* h1);
Thanks for the reply. I am sure that there are many complications
in a C/C++ interpreter and I am impressed that it can be
done so well. I was hoping though that the error that
finally is generated would also have a reference to the
string THF1 being unknown somehow. 2.11 was certainly an improvement
over 2.09. 2.11 has an execution time error about not finding
h1. 2.09 had no complaint at all, I just did not get anything
in my histogram.
Doug
> -----Original Message-----
> From: owner-starsofi-l@bnl.gov [mailto:owner-starsofi-l@bnl.gov]On
> Behalf Of Masaharu Goto
> Sent: Friday, September 11, 1998 6:59 AM
> To: roottalk@hpsalo.cern.ch
> Subject: RE:type bug in root
>
>
> Doug,
>
> >// THF1 does not exist, should be TH1F
> >int typebug(THF1* h1);
>
> I imagine you have spent long time to figure this out.
> I understand the problem and I'm sorry I can not fix it. This is
> not as simple
> as it looks. Consider following cases,
>
> typedef struct A { int a; } A;
> int a,b;
> int f(a* b); // error in C , variable f in C++
> int g(a); // error in C , variable f in C++
> int h(a) int a { return(a+3); } // function h in C, error in C++
> int x(A* b); // function x in boty C and C++
>
> In order to detect all errors, Cint needs to know whether the
> source is C or
> C++ in precise manner. This is a big problem because some user
> wants C++ as
> default and other user wants C as default. It is not possible to judge by
> file extension because .C is C++ in UNIX and C in Windows. Also
> .h can be
> anything in any platform.
>
> It is technically possible to solve this if I strictly determine C / C++
> distinction rule. The problem is that many users do not want it.
> Fixing this
> gives problems to many existing user applications.
>
> Masaharu Goto
>
>
>