durexforth VS zeptoforth

For the interested, there is a modern open source C64 Forth called DurexForth.

I know you said you wrote this in Kick Assembler, but I really feel like I should plug 64tass a fantastic cross-assembler. I've used it for my own reverse engineering projects. You might also want to look into DurexForth not because I think you should write this in Forth instead of ASM, but because it might help you to think about how to write things like a quest interpreter.

If you’re into this you may also like durex forth: https://github.com/jkotlinski/durexforth

A modern C64 variant, as well as CollapseOS which is a Forth based 8-bit OS: https://collapseos.org/

I initially implemented a rather Forth 2012-like wordlist system based on GET-ORDER, SET-ORDER, GET-CURRENT, SET-CURRENT, and WORDLIST for my Cortex-M Forth, zeptoforth. However, I ended up finding these quite cumbersome and error-prone to use in code that makes heavy use of wordlists to control the namespace, as is the case with zeptoforth. As a result I decided to completely remodel wordlists into a module system which, while internally based on those five words, is outwardly much more like the module systems found in other languages.

Multicore support for the RP2040 with zeptoforth has been in the works for a while but up until the last few days has not been mature enough for me to even consider including it in the devel branch of zeptoforth. However, now it has reached the point where I can run multiple tasks simultaneously on separate cores. I have a working test that blinks the LED on the Raspberry Pi Pico at two different rates in two different tasks, one on each core, while simultaneously writing an asterisk to the console once every second from the second core and having a usable REPL in the main task on the first core. It should be noted that even the Micropython does not do this, as it only allows two tasks, one per core, rather than allowing multiple tasks to run on each core separately.

I am the developer of zeptoforth, which is the main Forth I am using at the present. It supports the RP2040 (particularly the Raspberry Pi Pico, but it should work on other RP2040 boards), which I have been working with lately, and the STM32F407, STM32L476, and STM32F746 DISCOVERY boards. Note that it is not a desktop Forth; for that I would probably just recommend gforth.

zeptoforth 0.21.0 (https://github.com/tabemann/zeptoforth/releases/tag/v0.21.0) is now out, and introduces support for the RP2040 microcontroller (e.g. Raspberry Pi Pico, any RP2040 board with Winbond Quad SPI flash should work). Note that it comes in UF2 format, so one codes not need to solder pins for SWD onto one's Raspberry Pi Pico format to load it, and also the Makefile automatically generates UF2 files. Furthermore, when said UF2 file is first loaded, it erases flash above it up to the 1 MB mark, so one does not need to use OpenOCD (with SWD) or a special eraser UF2 file to clear old code out of flash.

The result of this is zeptoforth, which I have been developing for about a year and four months. One could say that it fits the same niche as Mecrisp-Stellaris, and it admittedly supports far fewer MCU's at the present. I could have just used that rather than bothered to implement my own Cortex-M Forth, but I wanted to create my own Forth environment which I could play with as I saw fit (e.g. adding preemptive multitasking and other multitasking supports such as locks and channels).