Re: Overloading vs. string functions, etc.

[Andreas Jaeger <aj at suse dot de>]

Mark Mitchell <mark@codesourcery.com> writes:

> >>>>> "Andreas" == Andreas Jaeger <aj@suse.de> writes:
> 
>     Andreas> So, what can glibc do and what should glibc not do under
>     Andreas> C++ compilation (and please explain why, I'd like to
> 
> I'm excited that you're willing to help with this.
> 
> How about, on GNU/Linux, just making our lives easy? :-)

That's what I'm considering but since I'm no C++ expert, I need help
from the experts.;-)

> Make <string.h> contain:
> 
>   #ifdef __cplusplus
>   namespace std {
>   #endif

But this would break with GCC 2.95, wouldn't it?  So it should be
(__GNUC_PREREQ just check that we use GCC with the supplied version):

#if defined __cplusplus && __GNUC_PREREQ (3,0)
namespace std {
#endif

...

#if defined __cplusplus && __GNUC_PREREQ (3,0)
}
#endif

> Also, add overloads for the variants of these functions that are not
> present in C?  And make the functions inlines that call
> `__builtin_foo' under C++?

I'm not sure what you really mean here.

> In other words, how about just making the C headers be valid C++
> headers, too?  This is Sun's approach, and it makes life much simpler
> for the rest of the C++ library.

I have to discuss this with the other glibc developers and this is
actually not yet implemented so might be too late for GCC 3.0. 

> Does this make sense?

Not really, I'm not familiar enough with C++ library design.

Andreas
-- 
 Andreas Jaeger
  SuSE Labs aj@suse.de
   private aj@arthur.inka.de
    http://www.suse.de/~aj