Progress Update

[crc]

It's been a while, so here's where I'm at:

1) repository split is complete:

VM: git clone git://github.com/crcx/ngaro.git
Image: git clone git://github.com/crcx/retro10.git
Library: git clone git://github.com/crcx/wheke.git

2) Toka is gone

We no longer need Toka for building the retroImage. Last night I
completed a working assembler/metacompiler and can now build the
retroImage with no additional tools.

3) VM

I've started work on porting the canvas support over to an SDL-backed
implementation. It's slow going, but the display, etc is working with
the standard image.

The initial image stuff is gone.

The C and .NET implementations gained a --shrink option which only
saves the used part of the heap during a save operation. This is
useful for making smaller images for sharing.

4) Core language

Back to three classes. (.word .macro and .data). The .compiler class
is now a loadable library (and does remap most of the .macro words to
use it).

5) Library

Expansions and cleanups are ongoing.

6) Easier installation

I've begun working on a shell script to download and install Retro on
unix-like systems: http://pastebin.com/f1c42e13c

-- crc

[Mat]

Hello,

I'm about to rewrite the AOT compiler of the extended VM for handling
of floating point, vector operations, parallel processing and general
simplification purposes. The rewrite will include backends for x86-64,
ARM and possibly PowerPC cpu's.

After that's finished I have plans for a retro interpreter which use
the compiler front end for code generation. There's also demand for
Lua, ML, Scheme and C compiler/interpreter combinations I don't know
how to find time for...

-Mat.

[Mat]

A little update:

I gave up using the C compiler for primitive generation because of bad
performance and some needed tricks of the generator which aren't ANSI
C conform.

Instead my current source code evolved into a very small compiler for
a functional colorforth dialect. My plan is to port my current Ngaro
VM on it after finishing the compiler. This offers the opportunity for
both meta compilation of Ngaro and Retro. I think about to create the
Ngaro sources for all ports from within the meta compiler.

Would this be a good idea ?

-Mat.