Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 2 Jun 1996 15:11:24 +0200 (MET DST)
From:      J Wunsch <j@uriah.heep.sax.de>
To:        freebsd-hackers@freebsd.org (FreeBSD hackers)
Cc:        fhackers@jraynard.demon.co.uk (James Raynard)
Subject:   Re: Two queries (libcompat.so and timedef())
Message-ID:  <199606021311.PAA01229@uriah.heep.sax.de>
In-Reply-To: <199606012029.UAA02321@jraynard.demon.co.uk> from James Raynard at "Jun 1, 96 08:29:48 pm"

next in thread | previous in thread | raw e-mail | index | archive | help
As James Raynard wrote:

> 1. ftok(). The prototype for this function is missing from
>    <sys/ipc.h>. No big deal, it's probably just an oversight which is

It should be in some system header (perhaps in <libcompat.h> if all
else fails), but not in some <sys/foo.h>.  The <sys/*> subhierarchy
is for declaring _kernel_ interfaces, not library interfaces.  It's
the primary source of include files for building the kernel.

I wonder where SysV did provide the prototype.  Perhaps nowhere. :)

>    On investigation, I have a libcompat.a and a libcompat_p.a on my
>    system, but no libcompat.so.

Intention.  It's not worth stuffing legacy functions into a shared
lib.

>    disk. Is this intentional? Also, is there any reason for ftok() to be
>    in libcompat, when all the other SysV IPC stuff is in libc?

It's a userland (i.e. library) thing, not a kernel one.  The other
SysV stuff in libc is simply the syscall wrappers only (as for all
other syscalls).

-- 
cheers, J"org

joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/ -- NIC: JW11-RIPE
Never trust an operating system you don't have sources for. ;-)



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199606021311.PAA01229>