Yukari Hafner
patreon
Status Update, January 2025
🕑 Added 2025-01-25 13:13:52 +0000 UTCOverview
Shinmera: released machine-state v1.2
Shinmera: released machine-measurements
Shinmera: released trial-examples v1.8
Shinmera: moved offices
Shinmera: updated cl-gltf, Trial to adjust for upstream changes in KHR_implicit_shapes and KHR_physics_rigid_bodies
Shinmera: improve gltf support in Trial to support multiple skins, rigs, and scenes per file
Shinmera: various other improvements to Trial
Scymtym: implement a gift-wrapping algorithm to improve convex-covering
Some Details
Hello! I hope you've all had a good holiday season and started well into the new year!
Since I can't help myself, I of course had a holiday project. Namely, I wanted to sculpt, rig, and animate a model of my avatar character, Yukari. The ultimate idea being that I'll be able to use that as a vtuber kind of thing on streams, and thus also regain the motivation to actually do streams.
The result of the modelling you've perhaps seen at the end of the 2024 In Review update. Since then I've also started integrating it into Trial, since I'm insane and don't want to rely on existing vtuber software. For that I had to make a bunch of adjustments to support multiple rigs, scenes, cameras, etc. per file in Trial, since I want to be able to set up all the scenes and so on in Blender.
Then scymtym and I worked some more on the collision detection stuff. The recent adjustments he's made to convex-covering have improved the behaviour considerably. However, I was still seeing weird issues in Weiss, which ultimately led me down the road of updating the Blender plugin for physics property exports we rely on. Because the spec of that is still in flux that changed, too, breaking cl-gltf and trial import. It's not the first time that's happened, but this time the fix was a lot more annoying to get right and I burned almost an entire week of work on it.
Anyway, we're back in working order now. In between that I also spent some time fixing up Kandria and correcting all of the bugs I could identify. There's still one minor thing I need to verify with regards to scrolling in Alloy, but that should happen within the coming week. So yes, it's long overdue, but a new Kandria release that fixes modio integration and Steam compatibility will finally be out.
In late December I also put out another brief job listing for someone to implement a Binary Space Partitioning tree in 3d-spaces. Our current approach with convex-decomposition is just too brittle, and I wanted something more robust and well-behaved for the static geometry. A BSP is a long-standing choice for that, but I really didn't want to have to implement yet another one of those structures myself, hence the listing. There were multiple fantastic applicants, and the lucky dice roll winner has already started working on the implementation! It should be done in time for the next monthly update, and I'm excited to see how it'll pan out for the collision accuracy and performance in Weiss.
The convex-covering/decomposition work won't be for nought, though. We still require this approach for decomposing dynamic objects, since constructing a BSP is quite computation intensive, and we can't update it on the fly like we have to for dynamic objects. So for that case we'll continue to rely on convex decomposition and the kd-tree for lookup.
Then in the second week of January the Swiss Game Hub (the coworking space I rent under) moved to its new location in Γrlikon. This is also where the European Lisp Symposium will take place in May! The move was quite exhausting, but I'm really happy with the new setup. It looks a lot better, and in general is a lot comfier and much roomier, too. The only real issue at the moment is that it still gets quite frosty in there, especially post move it was well below 20C.
In between all the messing about with collision stuff, I finally found some time to improve the general feel of the game. There's now a lot more blood being splattered about as you play and fight, making a real mess of the place. It's... surprisingly fun just with that alone, turning the environment into a disgusting mess is somehow very entertaining.
And that's about all for this month. Of course, there's also been some regular noodling with libraries like machine-state, manifolds, redist, etc, etc. I don't want to go on for too long, I'm quite exhausted from the 3 hours of motorcycle course I had this morning.
This coming week I'm taking a break together with my parents up in the alps. The weather report isn't looking too sunny, but getting away from my home always helps me relax a fair amount, so I'm really looking forward to it. I'm also looking forward to Charles returning to the SBCL/NX porting work in February. We hope to submit to ELS with that, too!
I hope you're all having a good time, and I'll see you in February!
β Love, Shinmera β€οΈ
