Re: [AD] No need for END_OF_MAIN (Windows)

[Evert Glebbeek <eglebbk@xxxxxxxxxx>]

> So close and yet so far. To bad there is no macro overloading, so we
> could have one for main() and other for main(int, char **)

Indeed. I think pretty much all ways to get rid of END_OF_MAIN() have been
explored several times over the past few years (including this one) and
none of them seems doable. Unfortunately.

> I search trough almac.h and alosx.h and only found this:
>
> > #ifndef ALLEGRO_NO_MAGIC_MAIN
> >    #define ALLEGRO_MAGIC_MAIN
> >    #define main _mangled_main
> >    #undef END_OF_MAIN
> >    #define END_OF_MAIN() void *_mangled_main_address =3D (void*)
> _mangled=
> _main;
> > #else
> >    #undef END_OF_MAIN
> >    #define END_OF_MAIN() void *_mangled_main_address;
> > #endif
>
> And as I said before didn't find where is _mangled_main_address used.

allegro/src/macosx/main.m, I guess. Note that the MacOS X problem is a
different one from Windows.

Evert