Re: [MiNT] mono

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

On Sun, Jan 31, 2010 at 2:34 PM, Standa Opichal <opichals@seznam.cz> wrote:
>> with the amount of open source kernels available, adding some of the
>> missing parts should not bee to much of a problem.
>
> True.
>
>> Getting the code to fit with FreeMiNT is another issue.
>
> This is the sentence you nailed it in. That's where I think it is faster to
> do it the other way around.
>
> With regards to getting the stuff into FreeMiNT it would either require
> a) modifying the parts for integration, which then complicates their
> maintenance.. downstream/upstream contributions..
> or
> b) building a compatibility layer for every outside code contributions to
> minimize the maintenance part, but that impacts performance
>
> Along the way it is about constantly finding issues on how to afix a modern,
> complex code base from outside to fit our aging, simpler interfaces.
> Providing simple interfaces (read the ones from GEM world) on top of more
> complex ones (the modern ones from outside) is less work and that's why I am
> suggesting to go in that direction.
>
> I'll rather stop here as I won't probably be contributing to either of thoseIt may inspire
> anyway (at least codewise)...
>
> Standa
>
It may inspire someone to take up MiNT Debian again, or try the NetBSD approach

What I would like to see (do) is Freemint kernel compile for x86 based
machines. GEM is designed to run at source level on either
achitecture, with motorola/intel formats built into various parts of
is systems, so GEM app sources need not be recoded. 99% of the command
line tools are *nix compatible or direct *nix ports, which is 99% x86
based (nowadays)

I might look into this after I have spent some time documenting the
MiNT System Calls, and get a better understanding of the way it does
things.

Paul
PS I have the same intention for a PPC as well...