Hi,

The macro 'stlsoft_sterile_class' does not seem to work for me. E.g. this code (I carried out everything to a single file, just to have it in one piece):
-----cut here-----
#define sterile_class(_cls)     \
  class __m__##_cls { \
   __m__##_cls(){}; \
   ~__m__##_cls(){}; friend class _cls; \
  }; \
  class _cls: public __m__##_cls


sterile_class(sterile_c) {
};

class BadClass : public sterile_c {
};

int main(int argc, char **argv) {
  sterile_c ss;
  BadClass s;
  return 0;
}
-----cut here-----
compiles and runs fine for me (I use mingw 3.2.3). In fact, I do not see why this macro *is supposed* to work.

Pavel

Argh!

This is one of those things that should have been removed.

Actually it used to work about 10 years ago, when most compilers were that way inclined.

I remember seeing it in the STLSoft header files last year and thinking I needed to remove it before someone discovered that it was useless. I will do so immediately.

Sorry for the waste of time (although I'm sure you've had cause to think about some inheritance stuff and all that, so hopefully not a complete waste <g>).

Cheers

Matthew

"pavel" <p_ro_te_ct_fromsp_amptolov@seznam.cz> wrote in message news:c6f570$2ssq$1@digitaldaemon.com...
> Hi,
>
> The macro 'stlsoft_sterile_class' does not seem to work for me. E.g. this code (I carried out everything to a single file, just to have it in one piece):
> -----cut here-----
> #define sterile_class(_cls)     \
>    class __m__##_cls { \
>     __m__##_cls(){}; \
>     ~__m__##_cls(){}; friend class _cls; \
>    }; \
>    class _cls: public __m__##_cls
>
>
> sterile_class(sterile_c) {
> };
>
> class BadClass : public sterile_c {
> };
>
> int main(int argc, char **argv) {
>    sterile_c ss;
>    BadClass s;
>    return 0;
> }
> -----cut here-----
> compiles and runs fine for me (I use mingw 3.2.3). In fact, I do not see why this macro *is supposed* to work.
>
> Pavel

They're now gone from the code that will be in 1.7.1 (or *another* beta; hopefully I'm pretty close to 1.7.1 release now).

Once again, sorry for the hassle.

"Matthew" <matthew.hat@stlsoft.dot.org> wrote in message news:c6frd1$sn9$1@digitaldaemon.com...
> Argh!
>
> This is one of those things that should have been removed.
>
> Actually it used to work about 10 years ago, when most compilers were that way inclined.
>
> I remember seeing it in the STLSoft header files last year and thinking I
needed
> to remove it before someone discovered that it was useless. I will do so immediately.
>
> Sorry for the waste of time (although I'm sure you've had cause to think about some inheritance stuff and all that, so hopefully not a complete waste <g>).
>
> Cheers
>
> Matthew
>
> "pavel" <p_ro_te_ct_fromsp_amptolov@seznam.cz> wrote in message news:c6f570$2ssq$1@digitaldaemon.com...
> > Hi,
> >
> > The macro 'stlsoft_sterile_class' does not seem to work for me. E.g. this code (I carried out everything to a single file, just to have it in one piece):
> > -----cut here-----
> > #define sterile_class(_cls)     \
> >    class __m__##_cls { \
> >     __m__##_cls(){}; \
> >     ~__m__##_cls(){}; friend class _cls; \
> >    }; \
> >    class _cls: public __m__##_cls
> >
> >
> > sterile_class(sterile_c) {
> > };
> >
> > class BadClass : public sterile_c {
> > };
> >
> > int main(int argc, char **argv) {
> >    sterile_c ss;
> >    BadClass s;
> >    return 0;
> > }
> > -----cut here-----
> > compiles and runs fine for me (I use mingw 3.2.3). In fact, I do not see why this macro *is supposed* to work.
> >
> > Pavel
>
>

Thanks Matthew!

No, it wasn't any waste. In fact, when I saw it first, I liked the idea of giving such "final classes in C++" and immediately decided to "beat it" with a template implementation -- only to find out that the template parameter could not be a friend (I wonder why though) and only then that the preprocessor implementation did not work either. This all took at most 30 minutes alltogether and amused me quite a bit.

Thanks again for all your hard work and finding time to answer
Pavel

Matthew wrote:
> Argh!
> 
> This is one of those things that should have been removed.
> 
> Actually it used to work about 10 years ago, when most compilers were that way
> inclined.
> 
> I remember seeing it in the STLSoft header files last year and thinking I needed
> to remove it before someone discovered that it was useless. I will do so
> immediately.
> 
> Sorry for the waste of time (although I'm sure you've had cause to think about
> some inheritance stuff and all that, so hopefully not a complete waste <g>).
> 
> Cheers
> 
> Matthew
> 
> "pavel" <p_ro_te_ct_fromsp_amptolov@seznam.cz> wrote in message
> news:c6f570$2ssq$1@digitaldaemon.com...
> 
>>Hi,
>>
>>The macro 'stlsoft_sterile_class' does not seem to work for me. E.g.
>>this code (I carried out everything to a single file, just to have it in
>>one piece):
>>-----cut here-----
>>#define sterile_class(_cls)     \
>>   class __m__##_cls { \
>>    __m__##_cls(){}; \
>>    ~__m__##_cls(){}; friend class _cls; \
>>   }; \
>>   class _cls: public __m__##_cls
>>
>>
>>sterile_class(sterile_c) {
>>};
>>
>>class BadClass : public sterile_c {
>>};
>>
>>int main(int argc, char **argv) {
>>   sterile_c ss;
>>   BadClass s;
>>   return 0;
>>}
>>-----cut here-----
>>compiles and runs fine for me (I use mingw 3.2.3). In fact, I do not see
>>why this macro *is supposed* to work.
>>
>>Pavel
> 
> 
>