Hello everyone. As I have posted in this thread, I am trying to get a BSD/UNIX userland on Linux, because I don't want GNU on my system, and I believe there are many people who want to see Linux free from the "*freedom* with GNU".

So I started this topic to see if there are people interested to help. I am planning to do a Linux from Scratch project where I will compile BSD (OpenBSD/NetBSD/FreeBSD; yet to be decided which userland to use) userland and then add those essential core components needed to make a functional, but minimal distro which can be further customized. Is anyone interested? (I need programmers/porters, but I myself will do the testing, etc..)

Whatever will be done here will be contributed to the whole open source community, and because I am a part of the AnthonOS team, I believe I will also work with them and talk if we can replace GNU with *BSD userland, and thus provide the Linux world with a desktop BSD/Linux distro.

I know this is not a very easy thing to acomplish, but is possible to achieve and everything starts out with an idea, so I hope there will be people interested to do this.

I also believe people will want to know why is there a need of a BSD userland on Linux, but I believe that many people would like to replace GNU for many different reasons, so we will compile list of reasons.
As for me, I think using the original UNIX/BSD tools (userland) is far better than using the GNU clones, because they are historic, far better developed and organized, and simply because they are originals, thus such system would have better POSIX compliance.

I would also take a look at Arch Linux as a base, but I think Linux From Scratch is better.

God bless!

EDIT: Actually hereis such a project which seems to be *dead*, but we (everyone who will participate) can try to contact with the dev and see if we can restart the project together.

But you realize, many many many things in many UNIXen now are GNU, such as /usr/bin/cc. Clang may work, but if you insist on historical accuracy, you'll be pretty much forced to use ansi2knr or something like that. And a lot of config scripts which assume Linux = GNU userland will break.

Basically, it'll be hell for users to use and compile software if the config scripts are broken.

I would go with clang.

Btw, what about ports system for compiling and installing packages?

Apart from the feasibility of the project itself, this is the first large obstacle.
You may break installation for many programs, because your system isn't GNU/Linux nor is BSD.

But you know, this fails also from the ideological point fo view.
If you were really striving for Unix historic accuracy / originality (what does it mean?!?), why would you couple an Unix original userland with a non-Unix kernel?
Provided you end up with a working system, that will never be an Unix system.

PS: there's not actual proof that modern BSDs are more POSIX compliant than GNU/Linux.
Yes, this is arguably true, because BSDs should be based on old BSDs man pages and documentation (this isn't really related to POSIX compliance, but whatever), but they don't contain any old BSD code anymore (did you know?).
Neither of them is certificated as UNIX by the Open Group, so we will never know.

Just my 2 cents, good luck.

This is what happens when you use a stock BSD userland on Linux. I did not install m4, strings, and cpp because well, I don't want to break gcc, or autoconf.

Why with a non-Unix kernel? Because there is much bigger hardware support for Linux.
Sadly, BSDs (and Solaris, and similar systems) don't have that support.

About the old BSD code, I thought that all BSDs we have now started with 4.4BSD code. And OpenBSD seems to be very faithful to that code.

It would still be better to finally have freedom from GNU and have the closest to original (if not original) userspace tools.

The mentality is, if it's not there, code it.

Mmm, is that the output of BSD ls on Linux?
It fails to get the correct size of files... cool.

BTW, which BSD flavour is it from and how did you get it to compile?
The only BSD source I got on my HDD is FreeBSD 8.0 (quite old, I know) and on Linux it fails to compile even the most basic utilities.


It might be easier to write missing drivers for FreeBSD than porting its userland to Linux.
Seriously, it makes more sense anyway.

True. But that isn't the only reason for a BSD userland on Linux. There are many.
I admit it is not an easy thing to do, but it is possible.

If this happens, it would be a great thing for the Linux community. We will have a choice. Otherwise, no matter how good does GNU or their ideology look, it is still a monopoly. They talk about freedom, but there is no freedom in reality. We are forced to use GNU no matter we like it or not. This is not freedom.

But still, BSD/Linux would be an interesting thing. I have contacted with the obase dev, and I will check his progress from time to time, so when he ports almost (if not) all OpenBSD's tools to Linux, I will try the LSF thing and I will post here.

Meanwhile if there are people interested in helping, please tell me.