| 1 |
On Fri, 2004-02-27 at 13:35, Chris Gianelloni wrote: |
| 2 |
> On Fri, 2004-02-27 at 05:25, John Nilsson wrote: |
| 3 |
> > I think you are wrong. ;) I think it *can* be argued sensibly. For these |
| 4 |
> > reasons. |
| 5 |
> > |
| 6 |
> > 1. Virtually all operating systems today ships with some GUI. |
| 7 |
> |
| 8 |
> ...and XFree86 4.3.x does not fit this bill? |
| 9 |
|
| 10 |
? |
| 11 |
|
| 12 |
> > 2. GNU (as in Gnu Public License) seems to regard the X Windows System |
| 13 |
> > as a core system component. |
| 14 |
> |
| 15 |
> X Windows System != XFree86 |
| 16 |
|
| 17 |
It is an implementation. You've got have an implementation in order to |
| 18 |
link against it, no? |
| 19 |
|
| 20 |
> > All you *need* for a linux based operating system is linux and a static |
| 21 |
> > binary called /sbin/init. Clearly the "Base system" referred to in GPL |
| 22 |
> > extends to more than that. |
| 23 |
> |
| 24 |
> Actually, you don't need init to get "Linux" at all. Following the same |
| 25 |
> path, you need nothing more than a kernel to have "Linux" at all. Now, |
| 26 |
> to have an actual working system, you generally need the kernel, an init |
| 27 |
> system, a few libraries (like glibc, uclib, etc) and a shell. |
| 28 |
|
| 29 |
Actually my initial statement was correct. If you do not modify the |
| 30 |
linux source all you need to have a RUNNING system is a staticly linked |
| 31 |
binary named /sbin/init. If the kernel can't find this it panics. When |
| 32 |
the init process dies the kernel dies. |
| 33 |
I did not say that this /sbin/init has to perform some actual function |
| 34 |
like the SysV init does, just be an infinit loop. Also it does not have |
| 35 |
to be a binary but the an interpreter would be needed. |
| 36 |
I once set up a firewall this way. The /sbin/init just configured |
| 37 |
iptables and put the log messages out to console. Nice one floppy |
| 38 |
firewall on a diskless 486. |
| 39 |
|
| 40 |
> > Even though I argue for compatibility, I still think it is correct to |
| 41 |
> > not ship XFree86. Mostly because Gentoo would and the OSS world would be |
| 42 |
> > far better of with a more "geekish" and open development of the X11 |
| 43 |
> > implementation. |
| 44 |
> |
| 45 |
> Gentoo is not going to stop shipping XFree86, as that would be asinine. |
| 46 |
> Rather we are simply not adding the NEW OFFENDING versions to portage. |
| 47 |
> There is a dramatic difference between the two. |
| 48 |
|
| 49 |
True |
| 50 |
|
| 51 |
-John |
Archives