Soupster [Lexaloffle Blog Feed]

Desktop window blending colour tables

Wed, 15 Oct 2025 20:41:05 UTC

When setting up a window for an app, you can set pixels to be transparent, so the background shows through. This allows us to have simple on/off transparency, but it does not allow us to blend colours and have translucent windows. My suggestion is to let the cart give the window a colour table. This would be applied before drawing the window and removed afterwards, and it would let the window blend with the background.

Get PNG pixel data?

Mon, 13 Oct 2025 19:11:59 UTC

In a recent update, fetch()ing a .png file will now return a paletter-fitted u8 userdata, instead of the i32 we had before. Unless I've missed something, this means it's now no loner possible to get the actual colours in the image? Maybe an extra argument to fetch could resolve this, as we still have the ability to pass true to userdata :convert to get the palette alongside a u8.

3D Picotron Screensaver

Wed, 21 May 2025 20:31:03 UTC

picotron_screensaver

by Soupster

[Click to Play]

Made a neat little screensaver that I thought I'd share. Smoothly transitions between some different scenes.

(Sped up a bit, but the GIF lags):

Mysterious bugs with pod()

Tue, 01 Apr 2025 23:56:31 UTC

I have encountered some strange behaviour when using pod() with keys that are tables.

This code snippet prints {} (it should print nil if it were to be consistent with other pod behaviour)

local a = {}
a[a] = 1

?pod(a)

This code snippet also prints {}

local a = {}
local b = {}
a[b] = 2

?pod(a)

But now, the strange part. This:

local a = {}
local b = {}
a[a] = 1
a[b] = 2

?pod(a)

causes an error, it reads simply attempt to compare two table values. No error location, just this!

Tile Building With Stress

Sat, 04 Jan 2025 12:04:00 UTC

tile_stress_test

by Soupster

[Click to Play]

Just a little test to see if the implementation works.

Controls:

LMB: Place
RMB: Pick tile

Left/Right arrows: Cycle through tiles

S: View stress
M: Change stress mode

Have fun!

Edit: Also, each tile has different properties, so you'll want to be careful what you build with.

BBS Bugs

Thu, 15 Aug 2024 22:37:11 UTC

Here is a list of bugs/annoyances in the BBS:

Important:

~~Code preview is broken, if you click on the code button under the pico-8 web player, it doesn't show the code. (It's the right size now, but still no code. Definitely an improvement!)~~ Fixed!

Minor:

The padding around the pico-8 web player is too big.
The featured buttons at the top of the BBS don't show the author, it just says "by ".

There are a couple more, but they're not really issues, so I've omitted them.

Picotron web player broken (possible fix?)

Tue, 13 Aug 2024 22:17:28 UTC

When clicking the play button on the Picotron web player, it stays as a grey screen. Checking the console shows an error, that it can't find https://www.lexaloffle.com/play/picotron_0101.js. After some digging, I believe it should be https://www.lexaloffle.com/play/picotron_0100h.js instead.

Edit: Wait, does this hint at Picotron 0.1.1?

FIXED!!!!

Picotron text field bug

Tue, 30 Jul 2024 20:11:31 UTC

Text gets "buffered" if no text editor is selected, and upon selecting it, it types the text.

How to reproduce:

Start Picotron
Type something
Open up filenav
Select the navbar
The text appears where you click instantly.

191-Character Maze Generator

Tue, 02 Jul 2024 01:49:54 UTC

This is a maze generator in 191 characters.

maze_gen_191

by Soupster

[Click to Play]

t={-2,2}function g(x,y)if(pget(x,y)<7)return
line(x,y)::l::r=rnd(t)if(rnd(2)>1)then
g(x+r,y)else
g(x,y+r)end
line(x,y)if(rnd(99)>1)goto l
end
cls(7)g(0,0)pal(7,8,1)pal(6,7,1)::_::flip()goto _

This is my first time doing this, let me know if it could be improved.

RGB Transition Effect

Fri, 07 Jun 2024 23:23:47 UTC

rgb_transition

by Soupster

[Click to Play]

This is a simple transition, that should be fairly easy to use. All the code is in rgb_transition.lua. Simply call rgb_transition_setup() when starting the transition, and then on each frame, after the game is drawn, call rgb_transition(). It takes one argument: Whether the transition should be fading in or out. True fades out. In this test cart, press X to reverse the transition.

3D Rendering with depth buffer!

Mon, 03 Jun 2024 19:07:58 UTC

3D Rendering with depth buffer!

by Soupster

[Click to Play]

This is mainly a tech demo for a 3d engine with a depth buffer. This is not finished, but I wanted to share it.

TODO:

Optimisations maybe?
BSP - currently model faces aren't sorted, so concave models will work weirdly. The depth buffer is only for interactions between models.
Maybe lower the resolution for better performance?
Some sort of culling?
Better use of depth buffer resolution, make it more linear maybe?

(The depth buffer works with the use of colour tables, in a few passes. If anyone wants more detail, I can provide it.)

Oh yeah, controls:
WASD: Move
QE: Up/Down (relative to look dir)
Arrows: Rotate camera

Z/C: Toggle depth rendering

Thanks @freds72 and @Maeve for advice on the memory issue, I hope it's gone once and for all! (Memory still rises rapidly, so I'll look into what could be causing that too)

f64 userdata doesn't save in pods

Thu, 23 May 2024 06:47:19 UTC

I've encountered a bug where vectors and f64 userdata doesn't get saved in the pod format.

This code:

local foo = vec(10, 10)

print(tostr(foo), 10, 10, 7)

print(pod(foo), 10, 20, 7)

print(tostr(unpod(pod(foo))), 10, 30, 7)

local bar = userdata("f64", 2)
bar.x = 10
bar.y = 10

print(tostr(bar), 10, 50, 7)

print(pod(bar), 10, 60, 7)

print(tostr(unpod(pod(bar))), 10, 70, 7)

prints:

(10.00000, 10.00000)
userdata("f64",2,"")
(0.00000, 0.00000)

(10.00000, 10.00000)
userdata("f64",2,"")
(0.00000, 0.00000)

You can see that pod() doesn't save the values for the userdata.

Pause when switching to picotron and pico-8

Wed, 22 May 2024 02:56:08 UTC

When switching to picotron or pico-8 by clicking on the icon when it's in fullscreen mode, I encounter a short freeze. Only about a second, but it's pretty annoying. @zep, do you know why this happens? Is it possible to remove?

create_diff missing? [Solved]

Tue, 30 Apr 2024 19:31:05 UTC

~~Hello. I was playing around with pods, and I noticed that create_diff isn't defined. here's the code (I'm 99% sure I didn't do something really stupid):~~

local a = {"a", "b", "c"}

local b = {"a", "c", "d"}

local pa = pod(a)
print(pa)

local pb = pod(b)
print(pb)

local pd = create_diff(pa, pb)
print(pd)

~~@zep was the function removed? it's in https://www.lexaloffle.com/dl/docs/picotron_pod.html#POD_Diffs, but maybe that's outdated.~~

EDIT: Never mind, it's called create_delta!

Conway's Game of Life

Mon, 22 Apr 2024 21:15:35 UTC

Conway's Game of Life

by Soupster

[Click to Play]

Here is Conway's Game of Life in Picotron, using the colour tables as a fast way to count all the pixels at once.
This version does not work in the web player, because input is not detected ( @zep pls fix!)

Controls:
Up: Increase simulation speed
Down: Decrease simulation speed
Space: Pause
F: Step one frame
C: Clear screen
R: Randomise screen
L-Click: Draw pixels
R-Click: Erase Pixels
L: Toggle large cursor

Basic code:

function _init()
	frame = 0

	-- Set up userdata so we can draw the screen to itself
	-- by using memcpy() to a memmapped region of memory.
	-- The userdata is now the contents of the previous frame,
	-- and it is now possible to modify the current frame.
	-- We can also call spr() with userdata
	-- to draw the previous frame to the current one

	screen = userdata("u8", 480, 270)
	memmap(0x30000, screen)
end

function set_col_table(new, current, col)
	poke(0x8000 + 64*new + current, col)
end

function _draw()
	frame += 1

	if frame == 1 then
		-- Randomise the screen for the first frame.
		for y=0,269 do
			for x=0,479 do
				if (rnd() < 0.2) pset(x, y, 7)
			end
		end

		-- Copy the current screen to the buffer, just for the first frame
		memcpy(0x30000, 0x10000, 0x20000)

	end

	--if (frame % 32 > 0) return

	cls()

	-- White cells drawn onto colour 0 will set the colour to 1, white onto 1 will be 2
	-- and so on. This counts the number of neighbouring cells very quickly
	for i=0, 7 do
		set_col_table(7, i, i+1)
	end

	-- Draw the screen 8 times in a ring, for each neighbour. The colour tables
	-- do the counting
	for y=-1,1 do
		for x=-1,1 do
			if (x!=0 or y!=0) spr(screen, x, y)
		end
	end

	-- Set up colour tables to turn the "counted" screen into the next frame
	-- Set every colour to black except for drawing black onto 3 (brought back alive)
	-- and 2 or 3 neighbours for alive (stay alive)
	for i=0, 9 do
		set_col_table(0, i, 0)
		set_col_table(7, i, 0)
	end
	set_col_table(0, 3, 7)
	set_col_table(7, 2, 7)
	set_col_table(7, 3, 7)

	-- Draw the screen to the "counted" version, with the rules set above
	spr(screen, 0, 0)

	-- Reset the draw state to make sure we have a predictable next frame
	reset()

	-- Copy the screen to the buffer for the next frame,
	-- before we pollute it with the FPS counter
	memcpy(0x30000, 0x10000, 0x20000)

	-- FPS counter
	if key("x") then
		rectfill(0, 0, 44, 8, 0)
		print("FPS: "..stat(7), 1, 1, 8)
	end

end