2024-01-19 19:04:59 <lispmacs[work]> hi all. For a work project, I have access to this twenty year old computer running QNX. Memory is a little tight, but it has a POSIX 1003.1 compatible Watcom C compiler already installed. I was wondering if I could get a useful forth installed on it.
2024-01-19 19:05:18 <Zarutian_iPad> probably
2024-01-19 19:05:19 <lispmacs[work]> useful, as far as, can interact with files and that sort of thing
2024-01-19 19:05:32 <lispmacs[work]> any thoughts on which forth system to try?
2024-01-19 19:07:20 <lispmacs[work]> I get the impression gforth is wrapped close to GCC, so not sure if that would be a good place to start
2024-01-19 19:07:38 <thrig       > gforth is kinda huge
2024-01-19 19:35:25 <lispmacs[work]> PForth (Portable Forth) - that sounds promising
2024-01-19 19:36:01 <lispmacs[work]> posix interface
2024-01-19 19:56:03 <KipIngram   > When I was exposed to it (a bit over 20 years ago) I found pforth quite easy to work with.  Built from source readily, was easy to modify, etc.
2024-01-19 19:57:00 <KipIngram   > It did have the  non-Forthy termios settings, though, and I had to fight a little to change that.
2024-01-19 20:07:57 <crc         > how tight are the memory limits?
2024-01-19 20:32:39 <crc         > a few more options on forth in c to try:
2024-01-19 20:32:44 <crc         > https://github.com/MitchBradley/cforth
2024-01-19 20:32:44 <crc         > https://gist.github.com/lbruder/10007431
2024-01-19 21:30:50 <lispmacs[work]> I gave pforth a try. A problem I'm running into is that pforth uses stdint.h and that doesn't seem to be provided by this old version of watcom
2024-01-19 21:32:54 <lispmacs[work]> looks like stdint.h is a C99 addition
2024-01-19 21:33:01 <lispmacs[work]> foiled!
2024-01-19 21:34:21 <lispmacs[work]> I suppose I could make one
2024-01-19 21:41:17 <lispmacs[work]> look like cforth depends on stdint also
2024-01-19 21:41:56 <lispmacs[work]> lbForth doesn't but also doesn't include an file i/o functionality
2024-01-19 21:42:24 <thrig       > forth is generally more DIY than batteries included
2024-01-19 21:42:38 <MrMobius    > stdint should be easy to replace
2024-01-19 21:43:12 <KipIngram   > I moved away from trying to do this stuff in C some years ago.  i find it all much more straightforward in assembly.
2024-01-19 21:43:37 <KipIngram   > There's no really good way to use the registers to full benefit otherwise.
2024-01-19 21:44:38 <lispmacs[work]> it looks like I'll need to come up with a stdint if I want to compile an modern C stuff
2024-01-19 21:44:45 <lispmacs[work]> seems reasonable enough
2024-01-19 21:45:26 <lispmacs[work]> probably most C programmers don't want just a vague notion of the size of their types
2024-01-19 21:46:06 <lispmacs[work]> C programming must have been interesting before that
2024-01-19 21:47:47 <KipIngram   > Isn't one of the points of structs to let you reference members of a set of things sharing a memory region without having to know their specific sizes?
2024-01-19 22:07:24 <MrMobius    > stdint is not that useful when everything is 8 or 16 bit
2024-01-19 22:12:08 <MrMobius    > I mean if architectures are 8 or 16 bit. you can still have 32 bit types.
2024-01-19 22:45:46 <zelgomer    > most codebases would define their own size typedefs. it's annoying to see people continue to do it in modern code when they could just include stdint.h