Hacker News new | past | comments | ask | show | jobs | submit login
I wrote a Forth implementation for x86 (2012) (davazp.net)
95 points by sea6ear on Feb 8, 2017 | hide | past | favorite | 11 comments



Man, the title plays it down a lot. I thought it was just an interpreter and nearly skipped over it instead of a Forth-machine OS, which is much neater.


Here the author. I loved working on this project, it was very satisfying building everything from scratch.

However, it would have been even better if somebody else would have joint.


Looks impressive! Are you planning to add support for networking?


I stopped working on the project like 4 years ago. I don't have plans to add anything else soon.

It would be nice though, and networking in particular is one of the more interesting things to implement I think. Maybe some day.


I have a joint for you buddy. :)


I'm loving the recent rise of posts about FORTH. It's always been a pleasure to play with, right the way back from the 80s.


Like implementing a Lisp, every programmer who likes programming languages should implement at least a Forth. However, this goes a bit further and is lovely.


The enthusiasm with which this was written almost matches the enthusiasm the author had about the project.


Is anyone able to run this on Mac/Darwin? I think it might need gcc:

  $ cc -fstrength-reduce -nostdinc -m32 -nostdlib -fno-builtin \
     -nostartfiles -nodefaultlibs -I. -ggdb -I.   -c -o boot.o boot.S
  clang: warning: optimization flag '-fstrength-reduce' is not supported
  clang: warning: argument unused during compilation: '-nostartfiles'
  boot.S:21:23: error: unexpected token in '.section' directive


You will need a gcc that can make ELF files since that is what qemu -kernel expects. I don't think homebrew has a gcc that can do this.


(2012)




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: