Subject: Re: new vfork(2) implementation
To: Erik E. Fair <email@example.com>
From: Bill Sommerfeld <firstname.lastname@example.org>
Date: 01/07/1998 14:13:41
> >Ask anyone involved with apollo's implementation of vfork() on
> >domain/os about what the letters "JLFU" stand for..
> I'd always heard that as "JLRU" from the insiders I knew ("SR10 - JLRU").
JLRU was the marketing version of the slogan. :-)
There were also JLFU T-shirts done up by folks in R&D.
Incidentally, the main reason why they did a "real" vfork() was that
apparently whichever version of the C shell they were using *required*
true vfork semantics on address-space sharing to behave properly.
(presumably it would have been broken on NetBSD, too..)
> It always sounded to me like they were trying to convince themselves...
They were. Many of the fine details of UNIX system call semantics are
difficult to "get right" when the location of the user/kernel split is
moved; if you come from a mindset which prefers giving the implementer
more freedom to play with things like that it can be a bit frustrating