Works fine. Let's see how it'll go from there.
So, I added the walk animations, collisions (with bump.lua) and an experimental transparency for objects way too close feature.
The latter is being extremely expensive on the CPU because Picotron shenanigans. I have to do something about that (and generally the whole overuse of sspr)
Addendum: I decided to only dither the biggest objects. No need for stuff like grass for now.
I threw together some coroutines to make random walkers on the island and eeeeeee it works!🥹 (well, they get stuck but i don't care!)
Nothing really fancy tonight, I'm cleaning up some bits and pondering as I want to introduce objects interactions.
In the mean time, I added also a top_down view with some debug utils to have a global vis of the scene and colliders.
Shaved roughly 1% of CPU on the current test scene by drawing batched rects for the houses' slices not in front and splitting the tree in two spirtes.
rectfill scales slightly better than sspr for the same amount of drawn pixels, so with some userdata sleight of hand, I can grab a few cycles back.
This is currently only visual but I've got the basics of interactive areas and detection.
I'd need to guess what's the object/entity the player wants to interact with but so far for the simpler cases it'll be good enough.
I split the project from a cartridge to a folder to edit it externally. Makes big refactors way easier.
I could clean up a good chunk of the code and made some optimizations to balance the extra tracking cost a bit everywhere.
I also rewrote the cylinder's draw arguments caching in a way that makes it reeally cheaper than even so I can almost pull camera position transitions live like that. The biggest cost sadly are the big objects, no luck there.
Hah! It doesn't look half bad. I'll probably need to make a whole mode out of it to test objects and placement (with extra "decals"/markings) but at least Blender's UV test texture works like a charm here.
Suffering from the engine dev blues. I want to make some tooling to help myself creating 3D objects (with slices /stacked sprites) and exploring stuff like bigger map systems.
At no point a good game idea is popping in my mind, it's only tech. (ノಠ益ಠ)ノ
Started an object viewer cart next to the game's cart (hard linking directly to the game cart's source files for rendering lol).
I hotswapped the map with a custom sprite I blit the size markers at runtime. Yes, I could theorically do splatting. Oh no, I'm giving myself cursed ideas.