Messages in this thread | | | Subject | Re: Header files and interfaces | Date | Fri, 26 Jun 1998 14:39:40 +0200 | From | Olaf Titz <> |
| |
> Actually, in this case net-tools really is almost completely to blame. It > used the presence or absence of OLD_SIOCSIFARP to decide whether or not > SIOCSIFARP actually referred to the "old" or "new" variant. This would have > gone wrong under libc5 as well if the OLD ioctl was deleted in some future > kernel.
What's wrong with keeping backwards compatibility? We've had the oldstat() stuff, we've had the "warning: uses broken flock() emulation" stuff, now we'll do the same with the cua* devices. All to make the transition to newer kernels easy.
Perhaps the old functions can be omitted from the second-to-next generation kernels so that they will give ENOSYS, but there's no point in deleting defines from the headers.
olaf
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu
| |