[Oberon] ARM:ETHO ?

[eas lab]

I used to think that gcc was mature/stable, but I recently had
bad experience trying to install Linux: Festival TextToSpeech
from Edenbourgh (sp) university, and that's where I REALLY needed
LEO to untangle the mess.

chris at gcjd.de-web.ws  wrote:-
> I would recommend aagainst this way unless you are experienced with gcc
> internals already. ;-)
> ...
> Several years ago I evaluated this way and unless things have changed
> very much since then gcc internals are a mess, documentation is soso and
> gcc is a huge and moving target not easy to handle if you have been
> exposed to Wirth style.

Yes moving target: the people who had succeeded in installing
festival, had the version of the OS & gcc that matched the version of
festival.   In contrast I installed LEO several years ago when using
Mandrake9, then when I move up to FC1, on a different partition,
I just linked to the old LEO and again when I moved up to Slak13
[to get better USB use], and LEO behaves the same under all
versions of Linux that I've used [which have X].

Peter Matthias' proposed project is probably the closest to
ARM:ETHO. I *MUST* have the TUI look-and-feel. But LEO's
easy acces to the dir-system instead of the continuation of the
flat-file-system which LNO catered for is necessary.  So I had
another look into LEO's docos:--
This is fom LEO's docos, where my comments are thus "=>"
=> Oberon.Desk  == Base Systems:
        x86  Linux      Tool.           (Linux Kernel >= 2.6 !!)
        x86  Solaris    Tool.
        sparc  Solaris  Tool.
        ppc  MacOS X    Tool.           (Darwin >= 6.2 !!)
        ppc  Linux      (no longer supported)  <--?
Does this REALLY support non-x86 system/s ?!
Can we find any actual code which PROVES that this is non-x86 ??
=> UnixOberon.Tool == <non-Desktop eqivalent tool>
...
        ppc  MacOS X    Tool.           (Darwin >= 6.2 !!)  <-- opened below ##
        ppc  Linux      (no longer supported)
## => Tools/ppc.Darwin.Tool == Oberon startup code ------------
        cd ppc/Darwin/startup <- search for
        make
-------------- these modules get loaded by Modules0 --------
=!=> OK, since we are now in LEO, which can access linux files, lets use
LEO to copy some of the file contents here [we've already used linux to
look at them]. But LEO is far superior for keeping notes.
-> System.Directory  ppc/Darwin/startup/*\d == <plenty> ...
ppc/Darwin/startup/CallOberon.s     17.01.2002 12:35:22    1555
...
ppc/Darwin/startup/Makefile     24.10.2003 23:43:07
-> EditTools.OpenUnix ppc/Darwin/startup/CallOberon.s ==
#
#       G. Feldmann     30.12.96
#
.text
        .align 2
.globl _CallOberon
_CallOberon:
        mflr    r0
        stw     r0,8(r1)
        stmw    r2,-120(r1)  <-- !! This is NON-x86 asm
        stwu    r1,-176(r1)
...
=!=> let's confirm what the corresponsing x86 asm looks like:-
There are NO src/x86/ *.s  files
OK, I need not have looked at ppc/Darwin since ppc/Linux is also non-x86,
but ppc/Darwin seems further removed from x86/Linux
and we want to know if we get do ARM/Linux.
The coding and documentation looks very orderly, and I'm optimistic
that we can use it to get to ETHO:ARM.Linux
Is Gunter Feldman available?

== Chris Glur.

It's very difficult for me to email since my African ISP is broken