[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Planning for the future (was: debootstrapping m68k-coldfire)

I would subscribe to an organizational mailing list if
you could find a host for That.

I also run debian on powerpc which as you may know
is officially "embedded" now.

I feel compelled to point out that embedded is not
a euphemism for slow. the fastest consumer and
engineering and research computers now are 2/3 embedded.
(specifically i refer to the ps3, ibm intellistation+, and
ibm blue gene -- all power arch, only intellistation considered
a desktop).

doesn't embedded usually mean specialized or/and customized.
perhaps dedicated.

the main function of 68k for me is to help me improve linux
skills. I am no systems guy and already served my time in admin.
but i interest in development from an experimental point of
view. debian sometimes feels too constrictive.

i have a really sore spot about having used 68k machines in their
original incarnations and later macos and the current linux is
terrribly inefficient. users of overpowered new machines don't
notice this basic situation. ten years ago linux on a powermac
was FAST. not today... at least not debian anyway.

BTW i installed release of netbsd for 68k which came out about
5months prior to debian-etch. it was supported for 68k as a
full release, in netbsd terms (they don't prebuild all the packages
there or test them on all arches -- but a user build some 1000 packages
about 6 months later for 68k - also they do have some splitting
and cooperating as you were describing). I think you might look
at it a little bit more. Netbsd is different than linux both tech and
politically -- in one way it is much more stable.

the only significant efficiency gain though i found in limited testing
was that xserver support lower bit depths which is faster and
less computational load.

On Wed, Mar 5, 2008 at 4:22 AM, Michael Casadevall
<sonicmctails@gmail.com> wrote:
> Well, the question is if both coldfire and m68k could be both hosted as
>  seperate ports remains to be seen. On one hand, we have ports like mips
>  and mipsel which only differ in the endian ordering. On the other hand,
>  Debian SuperH died because it would require four (?!) ports. I'm not an
>  expert in these matters (the only port I have seriously been involved with
>  before this is hurd-i386 ... yes, you can laugh now).
>  As for a mini-debian distro, we could create a Custom Debain
>  Distribition for m68k and such. Still, if we can provide a full Debian
>  archive in addition to a CDD, that would probably be ideal. Unfortantely,
>  with four buildds down, we're not in a good situation at the moment. In
>  addition, we're not going to be able to release with lenny (baring a
>  small^W major miracle).
>  Until we're at a state we can offically release with Debian again,
>  we should probably create an unoffical m68k Debian distro unless we want
>  to leave what few users we have left running etch-m68k (and considering
>  that we're looking at lenny+1 to re-release, thats a time period of quite
>  possibly years. I'm willing to host such a setup, and I got a friend I can
>  get into mirroring it (and a few other people here might). We can leave
>  our unstable here, but we should have stable releases (with updates)
>  hosted somewhere for users in the intern until we can meet release
>  criteria.
>  Here's what I purpose what we do (and if we can ever do the video
>  conferencing, I'd love to talk about this face to face).
>  - Figure out what we want to include in an embedded Custom Debian
>    Distribition. I can figure out quite a few things just sitting here
>    writing this emails. We should also offer this CDD for i686, ARM,
>    MIPS(el), and any other embedded architecture to generate interest.
>  - Bootstrap m68k-coldfire as a seperate port to interest coldfire
>    developers. If it proves fessiable to merge the two ports into one
>    binary port, we simply need to provide an upgrade path.
>  - If we ever get a CDD off the ground, look into offering uclibc as an
>    alternative since most embedded chips do not have an MMU
>  - Determine if we're going to break for Debian for the time being and
>    release as an unoffical distro for etch r4, lenny, and whatever else is
>    needed until we can merge back in.
>  There are probably a few things I'm missing, but that's all my brain can
>  muster at 7:22am.
>  Michael
>   On Wed, 5 Mar 2008, Geert Uytterhoeven wrote:
>  > Date: Wed, 5 Mar 2008 11:03:56 +0100 (CET)
>  > From: Geert Uytterhoeven <geert@linux-m68k.org>
>  > To: Roman Zippel <zippel@linux-m68k.org>
>  > Cc: debian-68k@lists.debian.org, Wouter Verhelst <wouter@debian.org>,
>  >     Michael Casadevall <sonicmctails@gmail.com>
>  > Subject: Re: debootstrapping m68k-coldfire
>  >
>  > On Tue, 4 Mar 2008, Roman Zippel wrote:
>  >> On Tuesday 4. March 2008, Wouter Verhelst wrote:
>  >>> Anyway, the problem isn't that bootstrapping coldfire is hard; I
>  can do
>  >>> that myself if needs be, and we'd have a working port within a few
>  >>> months[1]. The problem is that adding another port isn't going to be
>  >>> accepted by FTP masters: I don't recall who exactly, but an FTP master
>  >>> did tell me that a coldfire port in Debian would only be accepted
>  if it
>  >>> was either part of the m68k port, or replaced it entirely.
>  >>
>  >> IOW technical reasons have no value when politics are involved. :-(
>  >> If we force everything into a single port solely out of political reason, it
>  >> gets a whole lot less interesting...
>  >
>  > Why would a new separate port not be accepted? Because of disk and
>  > mirror space requirements, or because of the overhead of having an
>  > additional port (both in contrast to the (limited) audience of m68k and
>  > Coldfire)?
>  >
>  > If disk space and mirror space are the problems, perhaps there should be
>  > a `Debian light' with less packages? Several big and resource hungry
>  > packages will never be used on m68k anyway (and perhaps not even on the
>  > slightly faster Coldfire). I'm quite sure we could fit m68k and Coldfire
>  > versions of `Debian light' in the mirror space of a full Debian port.
>  >
>  > This could be useful for other ports as well, and lower the entry
>  > requirements for new ports (cfr. the several recently added architectures
>  > in the Linux kernel).
>  >
>  > If all else fails, maybe we should start thinking about OpenWRT/m68k or
>  > something like that...
>  >
>  > Gr{oetje,eeting}s,
>  >
>  >                                               Geert
>  >
>  > --
>  > Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
>  >
>  > In personal conversations with technical people, I call myself a hacker. But
>  > when I'm talking to journalists I just say "programmer" or something like that.
>  >                                                           -- Linus Torvalds
>  >
>  --
>  To UNSUBSCRIBE, email to debian-68k-REQUEST@lists.debian.org
>  with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

Reply to: