CW 9.3 compiler bug

Hi,
I'm a Windows CW user. I'm posting here since it doesn't seem there is much
traffic of codewarrior.windows NG. I'd like to know if it's possible to
submit bug report to MW. I know it's a bit non-sense to submit a bug report
for which you'll never see a correction, but I enjoyed so much developing
with CW that I would like to contribute anyway.

I post here a small example to reprodure the bug (tested with CW 9.3):

class base{};
template<class T>
class privata : public base{};

template<class T>
class classe_di_test : private privata<T*>
{
friend void f(classe_di_test<T>);
};

void g(base&);

template<class T>
void f(classe_di_test<T> t)
{
g(t);
}

void test()
{
classe_di_test<A> t;
f<>(t); //#1
//f<A>(t); #2
//(f<>)(t); #3
}

at #1 the compiler should complain that template function f<> hasn't the
access right to private parent class base. Instead it complains that
void f(classe_di_test<A>)
does not exist. #2 and #3 works fine, i.e. they give the correct compile
error.

Thanks

---
Gianluca Silvestri

In article <QHZKe.67010$fm.4404430@news4.tin.it>,
"Gianluca Silvestri" <clcppmod-poster DeleteThis @not.a.valid.address> wrote:

> I'm a Windows CW user. I'm posting here since it doesn't seem there is much
> traffic of codewarrior.windows NG. I'd like to know if it's possible to
> submit bug report to MW. I know it's a bit non-sense to submit a bug report
> for which you'll never see a correction, but I enjoyed so much developing
> with CW that I would like to contribute anyway.

See:
/Release Notes/Contacting Metrowerks/email Report
Forms/email_Bug_Report_Form.txt

> I post here a small example to reprodure the bug (tested with CW 9.3):

Try with CW 9.5?