Subject: Re: RFC: migration to a fully dynamically linked system
To: None <firstname.lastname@example.org>
From: Bill Studenmund <email@example.com>
Date: 12/29/2001 15:44:01
On Sat, 29 Dec 2001, Rick Kelly wrote:
> Bill Studenmund said:
> >I think it is more that the objections have boiled down to not wanting
> >change as opposed to specific features of the change. Also, I don't really
> >think the objections are the majority.
> It just boggles my mind that anyone would want to increase the level of
> convolution in NetBSD. So far we have stayed away from the mess of twisty
> little passages that is SVR4 and Linux. Also, one should consider how these
> changes will affect the performance of 68k and other very old systems.
First off, I have to ask, do you really think the tone of this paragraph
will contribute to useful discussion? Your word coice is demeaning of
differing views, never a good starting point.
Second off, I think there are things SVR4 and Linux have done right. Just
because these right things were done by SVR4 and Linux doesn't mean we
should not do them. I also think there are things they did wrong. I think
there are good things they tried to do, and didn't do so well. For the
latter, just because SVR4 or Linux did it wrong, doesn't mean we shouldn't
do it, just we shouldn't do it wrong like they did.
> >The fundamental point is that we want to be able to add locale support and
> >new authentication schemes to all(*) programs, even ones in /bin and
> >/sbin. We really need that to be able to move forward in a number of
> >directions that I gather the majority of the project folks (including
> >myself) want to move.
> All of the binaries in /bin and /sbin don't need locale support.
Uhm, how do you know?
Tracerouting rmkhome.com indicates frii.net is your ISP. The presence of
"denver" in the router names along the way and the fact that www.frii.com
indicates that frii.net, Front Range Internet Inc., is a regional ISP,
indicates you are probably located in Colorado, USA.
I admit I am in San Diego, CA, USA.
My point is that we are in English locales (USA English to be exact), the
locale built into the OS. We don't know if the binaries in /bin and /sbin
need locale support as the OS always has the one we need. To tell if they
need locale support, we should ask non-English speakers.
I suspect they are going to tell us that these programs do indeed need
locale support. For instance, ls is in /bin. How are we going to list the
contents of a directory containing say Japanese or Korean or Chinese
characters, without locale support? How are we going to correctly sort
output in these cases? Heh, even for French or Spanish or German, how are
we going to correctly sort names with acented characters correctly,
without locale support?
I don't think ls is the only such program. While not every program may
need locale support, some definitly do. It strikes me as much easier to
build all of them dynamic rather than just a few.
> How many megs of stuff does all this add to /? What happened to neat,
> complete operating systems that can fit easily on a 1 gig disk with room
> to spare? Or even a 250 meg disk?
You haven't been following the thread, have you. Luke reported numbers
where dynamic linking *reduces* the space in / that /bin and /sbin need.
:-) More so than the space needed by a crunchgen'd recovery tools dir.
Also, there are a lot of other things going on which are growing root.
Like IPV6 support (ifconfig grew, and we got ping6 just to name two). Or
the ability to fsck a number of different fs types.
For NetBSD's disk needs to not grow, we need to stop adding things. To do
that means we stop growing, which means we die. I don't want to see that
You do mention one good point, and that is about the speed-hit of dynamic
loading, especially on older machines. I think we need to look into that
too. But I feel that is a seperate issue. :-)