Quote: "so you ended up with a "wrapper"?"
Well, not really a wrapper, as it's a C API and I'm working in C++, so I can just use it directly. I'd only need to wrap it if I wanted to expose it to something else, like writing a plugin or something.
Quote: "Would your program also run in Android or other than Windows in this way?"
It runs fine on Windows and Android. I've not compiled on the other platforms since my changes yet, but I don't see why not. it's all written in ANSI C so it should be fine.
Quote: "converting Spine to Blender would need a python-thingly in Blender"
Yeah but I mean, are we talking about a Python script that already exists, or do you mean I could write one?
Quote: " Since I have Spine Pro, I also can do Mesh-deformations, and some otherthings, which are not played correctly in AppGameKit Tier 1 Basic"
Yeah the support seems really limited at the moment. Even though I'm only on Spine Essential, I'm still missing features like events etc.
Quote: "I am also thinking about switching from Tier 1 to Tier 2, but a lot things are not work well in Tier 2"
What things are you having trouble with? I've been using Tier 2 for several years now and haven't had too many problems, although I'll admit setting up the templates can be a bit of a pain.
Quote: "maybe Tier 2 is later easier for port and combine with other C++ libraries?"
Yeah this is one of the (many) big advantages of Tier 2. I integrated FreeType 2 a couple of years before AppGameKit got truetype font support, and although I've moved over to using AGK's implementation again now, my current project is still making use of other 3rd party libraries like Lua and GLM (and now Spine), as well as the various different platforms' threading APIs. I try to keep this to a minimum, because it adds to my workload, and one of the big advantages of using AppGameKit over a custom engine for me is its out of the box cross platform support. That said though, being able to bring in third party libraries when you need them can be a lifesaver (as in this case - I would be completely stuck if I were working in Tier 1).
FWIW, moving over to using the Spine runtime directly has made a huge different to my load time. The initial cost of parsing the JSON file is exactly the same, but now I only have to do it once instead of for every character, and given that I can easily have ~100 characters in the game, this makes a really big difference.