> -----BEGIN PGP SIGNED MESSAGE-----
>
> adding an include <linux/socket.h> just before the include <sys/un.h>
> appears to have allowed it to compile. I will test it in the morning when
> I get back into work.
>
> I did verify that what bind/src/bin/named/pathnames.c includes is sys/un.h
> not linux/un.h
Libc5 strikes again. Sigh... At least use sys/socket.h instead of
linux/socket.h - on libc5 it should give the same effect and on other
systems it will be harmless. On libc6 (aka glibc) sys/un.h doesn't bring
linux/un.h and doesn't require socket.h. So more or less clean way to deal
with it would be
#include <sys/socket.h>
#include <sys/un.h>
Frankly, I don't know who's at fault here. I would say that libc5
should include socket.h from un.h, but then there is a question - WTF
would anybody do with un.h without socket.h? From what I see pathnames.c
per se doesn't need uh.h - it's for one of the .h files included from
there. And all stuff used in pathnames.c is a couple of string constants.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/