Subject: Re: vfork manpage update
To: VaX#n8 <vax@linkdead.paranoia.com>
From: Greg A. Woods <woods@kuma.web.net>
List: current-users
Date: 08/25/1996 13:32:16
[ On Sat, August 24, 1996 at 08:58:50 (-0500), vax@linkdead.paranoia.com wrote: ]
> Subject: vfork manpage update
>
> The vfork manpage is out-of-date. It implies that vfork has the old
> semantics still. It alludes to the new semantics in the BUGS section.
> A brief look at sys/kern_fork.c seems to suggest it's already here
> (as does D&I/4.4BSDOS).
This is really off the track, but FYI....
Solaris-2.x has re-instated the original semantics of vfork() -- which
actually has some advantages to them with their kernel-level
light-weight threads. Some widely used programs, such as GNU Emacs,
already assumes in bad ways (i.e. with signals, etc.) that vfork() does
have the exact same semantics as fork(), and now it needs to be taught
the old ways again! ;-)
--
Greg A. Woods
+1 416 443-1734 VE3TCP robohack!woods
Planix, Inc. <woods@planix.com>; Secrets Of The Weird <woods@weird.com>