Re: "undef. ref. to __va_copy"??

[Drazen Kacar <dave srce hr>]

James Henstridge wrote:
> Because there isn't a standard way to *copy* a va_list structure.  On some
> machines it is a pointer, and on others it is an array.  I guess egcs
> offers more helper routines to do these sort of things than the older
> versions of gcc.
> 
> I suppose the rationale is that if the functionality is there, you may as
> well use it.

> > You mean that glib is compiler dependant? Just like C++ libraries?
> > Why doesn't it just #include <varargs.h> if it exists on a given
> > platform?

Yeah, yeah, yeah. But I have GTK compiled with cc (on Solaris) and an
application which happens to use varargs with one gtk function.
When I'm compiling, both compilers (cc and gcc) warn that va_start and
va_end are redefined and it doesn't work with either. Guess why.

The functionality is there, but I can't use it any more. I won't even
try with the third compiler (which is not egcs, BTW). Another GTK miracle...

-- 
 .-.   .-.    Life is a sexually transmitted disease.
(_  \ /  _)
     |        dave@srce.hr
     |        dave@fly.cc.fer.hr