New warnings: INT redefined

Jakub Jelinek jakub@redhat.com
Tue Feb 5 05:12:00 GMT 2002


On Tue, Feb 05, 2002 at 02:04:28PM +0100, Andreas Jaeger wrote:
> 
> Using INT in libc-symbols was not the best idea, we now get a lot of
> warnings since INT was used previously:
> 
> ../sysdeps/generic/strtol.c:62:1: warning: "INT" redefined
> ../include/libc-symbols.h:355:1: warning: this is the location of the previous definition
> ../sysdeps/generic/strtol.c:64:1: warning: "INT" redefined
> ../include/libc-symbols.h:355:1: warning: this is the location of the previous definition
> tst-printf.c:77:1: warning: "INT" redefined
> ../include/libc-symbols.h:358:1: warning: this is the location of the previous definition
> 
> How should we solve this issue?  I would suggest to use INTNAME
> instead of INT in libc-symbols.h.
> 
> What do you think?  Shall I send a patch?

I'd say it would be much cleaner if elf/Makefile added -DRTLD_OBJECT
to all ld.so only objects and if ldsodefs.h and maybe other headers used:

extern void _dl_debug_printf (const char *fmt, ...)
     __attribute__ ((__format__ (__printf__, 1, 2)));
extern void _dl_debug_printf_internal (const char *fmt, ...)
     __attribute__ ((__format__ (__printf__, 1, 2)));
#ifdef RTLD_OBJECT
_dl_debug_printf(...) _dl_debug_printf_internal(__VA_ARGS__)
#endif

Then, only INTDEF would be needed...

	Jakub



More information about the Libc-hacker mailing list