Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 21 Sep 1995 14:06:54 -0700
From:      patl@asimov.volant.org
To:        chuckr@eng.umd.edu, gryphon@healer.com, kelly@fsl.noaa.gov
Cc:        asami@cs.berkeley.edu, hackers@freebsd.org, julian@ref.tfs.com, ports@freebsd.org, terry@lambert.org
Subject:   Re: ports startup scripts
Message-ID:  <9509212106.AA22801@asimov.volant.org>

next in thread | raw e-mail | index | archive | help
|>  Pat writes:
|>  > Coranth Gryphon <gryphon@healer.com> wrote:
|>  > +>  When are you going to start a daemon in more than one place?
|>  > +>  Or set a global environment variable, then change it later?
|>  > +>  
|>  > +>  Define it as "Run Level N" includes all "Run Level 0..N-1". Simple.
|>  
|>  > If it really were that simple, why didn't SVr4 do it that way?
|>  
|>  You're asking why an OS that I think has a lousy implementation
|>  of run levels didn't do it right in the first place? :-)
|>  
|>  Answer: I don't know why they didn't do it right. That doesn't
|>  stop us from doing it right, does it?

Perhaps I was being too subtle.  I meant to point out that it might
-NOT- be as simple as it looks at first glance.  We should find out
why they did it the way they did before we reject it.  And we should
investigate any existing variations.  (Does Solaris 2.4 do it the same
way as SVr4?  Does HP-UX?  Does...)


|>  > But differences just because "we aren't System V" only hurt us.  We
|>  
|>  Granted. But keeping it identical just because they did like that is
|>  just as bad. Takes the best parts, the ideas that work. Keep anything
|>  that does not make a difference the same (to satisfy the "no gratuitous
|>  changes" camp), but be willing to change what we don't like for our sake.

Right.

|>  > don't want people refusing to run FreeBSD because it is too different
|>  > from the other unixes, (without significant advantage) do we?
|>  
|>  Then keep it only BSD and stop trying to System-V-ize it.

Nope, won't work.  That just means that other unixes will continue to
evolve away from us, and increase the degree of differences.

|>  > +>  8-10 unix flavors that I work with. So which one are you going
|>  > +>  to clone?
|>  
|>  >We have two reasonable choices:
|>  >    1.	Whichever one we feel is technically superior.
|>  >    2.	The one with the biggest market presence (Solaris2).
|>  
|>  Neither. Build our own implementation, the way FreeBSD wants it,
|>  that conforms to the basic framework.

Right.  I was responding within the implicit assumption that we were
going to clone something.  I should have made it clear that cloning
wasn't the only (or possibly even the best) solution.

|>  If I want to use Solaris, I'll use Solaris. I don't want FreeBSD
|>  to just become a cheap Solaris clone.

Worse things could happen.  But I agree - FreeBSD shouldn't become
a cheap -anything- clone.  On the other hand, Solaris has done a
lot of things pretty well, and a lot of unix users are familiar with
it.  If their solutions fit our problems, and we can't come up with
a significantly better solution, it is reasonable to consider copying
the Solaris solution.

|>  > The current /usr/share/skel is a fine location.
|>  
|>  Except when you want to change default dot files. Then have
|>  to redo it after each install.

That's a major problem with our install process.  One of the nice
things about the Solaris pkg_add system is that the package can
have per-file actions for how to deal with files that have changed
since the previous install.  (The most common choices are: 1. Rename
the changed version and install the new one.  2. Install the new
one under a different name.  3. Just leave the old one in place.
But the install script can do anything you like.)



-Pat



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?9509212106.AA22801>