[lkml]   [1996]   [Jun]   [25]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: No Distribution is 2.0.0 Current
On Mon, 24 Jun 1996, hab wrote:

> Kevin M Bealer wrote:
> >
> > On Mon, 24 Jun 1996, hab wrote:
> >
> > > In my opinion no Distribution is Linux 2.0.0 current.
> > > To me to be 2.0.0 current all free packages must be compiled
> > > against the Linux 2.0.0 source tree, with the same libc or later than
> > > the developers use. It must also include the kernel support utilities
> > > that have been patched to work with 2.0.0 and all support tools need
> > > to be updated to support all the standard features of 2.0.0.
> > >
> > (clip)
> >
> > Just for reference, Debian does not necessarily compile against the kernel
> > anyway. Debian ships libc with it's own version of the kernel headers, from
> > a "known stable" kernel. If you want the actual kernel headers from the
> What are known stable kernels? 1.2.13 was the last one I know about prior to
> 2.0.0.

Kernels with no known bugs in the kernel headers. If 1.3.77's headers had
no bugs corrected until say 1.3.80, (this is entirely hypothetical) then the
1.3.77 headers would be less likely to cause problems when compiling non
kernel related utilities.

> > current kernel (only really necessary for programs which are very kernel
> > dependant, ie userfs), you can make symbolic links, etc and do this.
> >
> I totally agree that a stable platform is required. Thus my argument for don't
> get to quick in calling a system 2.0.0 current. 2.0.0 and its patches will be around
> for a year or better as a known Stable Kernel. Thus the plea to make it the true
> base for stable distributions.

I think the whole policy with libc versions vs. kernel versions was to
allow people with "otherwise stable" systems to use development kernels.
With the "stable tree" I believe the policy has little effect.

> > This allows people who use the "development" distribution of Debian to try
> > out new kernels with less risk of breaking compilation of other programs.
> >
> By Debians own policies 2.0.0 should be out for 3 Months before becoming part of
> a stable release? Or are their unknown exceptions to policies?

No... where are you getting 3 months from?

> > By your definition then, Debian may never be current, even if it was
> > completely recompiled with 2.0.0 installed.
> >
> I don't understand. I advocated finding out from a package developer what
> version of the package he considered debugged stable with 2.0.0. This
> should then be compiled against a common base with all other packages.
> The whole point is to get away from random bugs and incompatabilities
> caused by ramdom histories. Why should I need multiple versions of libs
> to keep all programs from one distribution running. I would like a known
> stable base, so I can easily isolate my tools from my research.
> > BTW don't go to debian to argue out this policy, it has already been
> > debated to death on those lists.
> Thanks for the warning. I guess Debian will never be the choice for a
> serious Computer Professional.

I see you capitalize "Computer Professional"... is it a name brand?

> >
> > __kmb203@psu.edu_________________________Debian__1.1___Linux__2.0.0___
> > It is difficult to produce a television documentary that is both
> > incisive and probing when every twelve minutes one is interrupted by
> > twelve dancing rabbits singing about toilet paper.
> > -- Rod Serling
> If part of your disagreement was my Use of Libc-5.3.12 which I couldn't find in the
> Development tree of Debian. I guess Debian will never be secure enough for my use as
> I will not knowingly use an inferior product.
> Hubert Bahr

I think by the time everything is recompiled for 2.0.0, we will have 2.0.3
or 2.0.7. Stabilization takes time, you can't set a standard that will
absolutely guarantee stability, or any other such quality. For many many
users, a security bug in 5.2.18 is a better tradeoff than the unknown
problems in a less tested, newer libc... anyone who needs that security
already knows which libc's to get. Paranoia is good; it is necessary. It
is not for everyone-at-any-cost.

It is difficult to produce a television documentary that is both
incisive and probing when every twelve minutes one is interrupted by
twelve dancing rabbits singing about toilet paper.
-- Rod Serling

 \ /
  Last update: 2005-03-22 13:37    [W:0.046 / U:2.812 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site