Re: [MiNT] Gentoo FreeMiNT

[Paul Wratt <paul.wratt@gmail.com>]

On Tue, Feb 2, 2010 at 3:06 PM, Mark Duckworth
<mduckworth@atari-source.org> wrote:
>
> Under intense testing I found the ccache did some very bad things and
> introduced corruption.  I didn't trace down the source.  Not recommended at
> this time by me.  ccache running on a host with distcc however may work just
> fine.
>
> Thanks,
> Mark
>

running ccache host side makes more sense when using ARAnyM, and the
speed increase is probably noticable to.

It would be good if the ccache failures native side could be
documented, even if its just console dumps atm, it gives someone else
the the chance to look for the problem first, as I imagine with CF it
would be a practical solution

For Peter S benifit, Gentoo Linux (being gneration 2 linux) was
initiated to fill the void created by linux running on new systems
(i586/i686), and the number of people still using and installing on
the sub par systems (486DX,386SX, 286). The idea was to provide the OS
in source form, compiled at installation, generating binaries that
excluded code not usable by your particular setup, thereby free some
of the bloat and resource consumed by tradition RPM/DEB route.

This is a similar situation as we currently have with ports for mint,
there is often a ton of code not useful to MiNT/TOS, and often the
code does not use the hardware in the best way (because this is not a
concern with modern machines).

The question begs (being a good idea and all, especially for ports)
how does this affect MiNT/TOS. The answer appears to be that in
reality only CPU is different between systems, and so the ability of
ebuild to install binaries is a perfect match for MiNT/TOS based
platforms.

One needs to remember that RPM was revolutionary in its day, and as
Mark can attest, it greatly simplifies the process of installing and
compiles apps something which can be an infinitly complex task when
applied to a whole OS

I think it is a good thing, to have more than one distro available,
and Mark is doing the first step in updating the RPM distro to new
standards (compiled with GCC 4.4.x).

Because of RPM's age, along with DEB and now ebuild, there are cross
build script for most packages, which in itself will hugely benefit
any modern posix kernel, which is perfect for MiNT/TOS.

After the initial libs and apps, which must be done for any distro,
the concentration on packages can diverge (between distros), allowing
for more packages to become available faster, which is only a good
thing.

I expect the the Debian GNU/MiNT will be updated again soon, maybe a
year or two away, but it will happen.

The one thing people contributing on this list seem to forget, is that
most people who use MiNT/TOS/GEM are not advanced users in the world
of distros, files systems, etc..

Regular (technical) posters need to remember (when replying) that most
people now a TOS filesystem. They usually use file tree between 2 & 3
levels deep at max. mostly things are simple, and if you have
problems, you boot on floppy to fix your AUTO folder, thats it.. MiNT
and distros are more complex than that, and appear seriously so if you
have never seen or used Linux, which many have not..

Also you need to remember, besides all the work that has been going on
over the years, and especially now, the MiNT and distro stability is
sub par, so you must forgive users of those systems, who are unable to
contribute to the core/kernel, for the question of "why another
distro" when the current ones are less than satisfactory, it looked
good, had all the tools, but would break regularly unless you had
built up a system of workarounds, which were not available to others
or as part of a distro, unless you were in the know

I dare say tho, with the volume of patches and distro work, as well as
other app and site work, going on at the moment, that within 6 months,
a sad system will be a thing of the past

Hope that eases a few mind as far as contention on both sides go..

Paul