2024-12-05 07:45:14 <lilo_booter > veltas: https://gitlab.com/lilo_booter/plotify <- this is the plotify project
2024-12-05 07:47:33 <lilo_booter > i haven't posted the image rendering yet - just cleaning it up (should be somewhat less embarassing by this evening :)) - basically, i've added the means to write ppm image files
2024-12-05 07:48:10 <lilo_booter > xentrac: and yes, well spotted - the image was indeed upside down :)
2024-12-05 07:52:50 <lilo_booter > fwiw, i'm totally cheating on the presentation - basically, i create a PPM file, and then open it using xdg-open which associates PPM images with "eye of gnome" (eog) - that shows the image, but also does a watch for changes, so each time i save the file, it refreshes (well, some of the time :))
2024-12-05 07:54:06 <lilo_booter > i think i want to write the image to a udp socket and have another process read/show what's coming through the pipe instead
2024-12-05 09:03:14 <lilo_booter > just had time to clean that up before starting work - ppm stuff pushed - default is terminal - ppm_init and term_init can be used to toggle between them (redrawing graphs along the way)
2024-12-05 09:03:23 <lilo_booter > 
ACTION does little happy dance

2024-12-05 09:09:59 <lilo_booter > relying on the existence of xdg-open and it associating ppm files with watch like functionality is a major hack of course :D, but otherwise i'd be interested in a critique of what's there from anyone (i have quite a number of gforth-isms which will make the whole thing quite unportable i suspect)
2024-12-05 10:19:34 <lilo_booter > sigh - that should have been ppm-init and term-init to switch modes
2024-12-05 10:28:33 <veltas      > Usually a bad idea to write 'portable' Forth anyway
2024-12-05 10:32:31 <lilo_booter > yeah - i kinda get that impression...
2024-12-05 19:46:09 <lilo_booter > couldn't be bothered trying to come up with a forth implementation to replace the xdg-open/eog usage, so knocked up a quick and dirty python script (yes, i could use sdl2 but i just couldn't be bothered with the event bindings necessary to implement the file watch functionality today)
2024-12-05 21:06:25 <KipIngram   > I take the most pleasure from those low level things.
2024-12-05 21:07:29 <KipIngram   > I went to a fair bit of trouble on my last couple of implementations to use nothing but actual system calls for the unavoidable hardware interfacing.
2024-12-05 21:08:06 <KipIngram   > It was originallly done on a Mac, and Apple apparently doesn't like people using system calls - they have a library that sits on top of that that they prefer you use, but I didn't.
2024-12-05 21:08:18 <KipIngram   > Which definitely made it easier to port to Linux later.
2024-12-05 21:08:32 <KipIngram   > Still took a bit of work, because the termios "magic numbers" are different.
2024-12-05 21:08:42 <KipIngram   > And really not well-documented at all.
2024-12-05 21:09:36 <KipIngram   > There are a lot of docs out there, but they all use the C symbol rather than the numerical value.  I wound up having to write a C program that just printed a list of symbol values - whenever I needed a new one I added it to the list and re-ran it.
2024-12-05 21:11:06 <xentrac     > yeah, I've done that pretty often
2024-12-05 21:11:17 <xentrac     > on Linux they are of course documented
2024-12-05 21:12:15 <xentrac     > I imagine the Darwin project also documents theirs
2024-12-05 21:46:53 <KipIngram   > Yeah, probably just I didn't know where to look.
2024-12-05 21:47:29 <KipIngram   > That stuff's only needed briefly at startup and again at exit.
2024-12-05 21:47:47 <KipIngram   > Just to make it possible to put the ok prompt at the end of the line instead of the beginning of the next line.
2024-12-05 21:48:12 <KipIngram   > Well, and also to give me access to every keystroke, instead of just lines full of characters.
2024-12-05 21:48:34 <KipIngram   > I turned off most of the control key handline - KEY just returns a different code for those in my systems.
2024-12-05 21:49:27 <KipIngram   > I think ctrl-M was the only one that was "stuck," so to speak, since it's the same code Enter returns.  Or maybe it was ctrl-J.  Whichever one Enter does.
2024-12-05 21:49:55 <KipIngram   > Yeah, ctrl-m.
2024-12-05 21:50:53 <KipIngram   > And I have my terminal program set so ctrl-c and ctrl-v are copy and paste, and that overrides.
2024-12-05 21:51:06 <KipIngram   > It seems common these days for it to be shift-ctrl-c/v, but I just don't like that - the habit is too deep.
2024-12-05 22:10:54 <xentrac     > yeah, my habit breaks every time I alt-tab to firefox, where shift-ctrl-c opens the HTML inspector