> That isn't a good solution. The breakage is in kfreebsd-kernel-headers, which
> expect the __CONCAT definition they obtain from <sys/cdefs.h> to be sane (and
> __CONCAT is not really a bsdism; not currently at least).
Unfortunately, the old and new behaviour will be
different when applied on macros:
#define __CONCAT1(x,y) x ## y
#define __CONCATN(x,y) __CONCAT1(x,y)
#define MYFUNC myfunc
/* old one */
/* new one */
> We could try that.. But I don't know well the rationale behind this bug.
> Would you like to file it?