Shouldn't fixincludes remove assert.h instead of trying to fix it?

Manfred Hollstein manfred@s-direktnet.de
Mon Mar 16 03:15:00 GMT 1998


fixincludes is supposed to provide fixed copies of the system specific 
include files to be used by gcc only. Since gcc installs its own
version of <assert.h>, I don't understand why fixincludes tries to fix 
a probably broken assert.h?

# assert.h and sys/mman.h on HP/UX are not C++ ready, even though
# NO_IMPLICIT_EXTERN_C is defined on HP/UX.

and

# check for broken assert.h that needs stdio.h or stdlib.h

Does anybody object, if fixincludes would remove this particular file
instead of trying to fix it? If not, I'll be going to submit a small
patch.

I'm asking this, because my `fixed' assert.h on m68k-motorola-sysv
actually causes several errors during make check:

#ifdef __cplusplus
extern "C" {
#endif
/*	@(#)head/assert.h	50.1	 */
#ifdef NDEBUG
#define assert(EX)
#else
extern void _assert();
#define assert(EX) if (EX) ; else _assert("EX", __FILE__, __LINE__)
#endif
#ifdef __cplusplus
}
#endif


Thanks

manfred



More information about the Gcc-bugs mailing list