[Looking for Contributors] I am planning my new language that transpiles to SRB2 Lua.

I'm planning to create a programming language that transpiles to SRB2 Lua and could improve mod development for SRB2.

---

---

The name is temporary. I'm still looking for a better one.

The long-term goals are:

• Have bytecode that can be transpiled into SRB2 Lua, albeit with obfuscated names.
• Support static typing and generics.
• Resolve modules by opening other .ytl files, performing semantic analysis on their exports, and reporting an error when code uses something that is not exported.
• Support static typing, with runtime asserts automatically inserted into the transpiled output.
• Include a virtual machine. (planned)
• Support plugins. (planned)
• Provide safe member navigation: optional with ?. and required with !..
• Preserve types, interfaces, enums, and classes instead of erasing them.
• Automatically generate LuaLS annotations.
• Support async using coroutines.
• Support dictionaries, tuples, and arrays.
• Support with, switch, struct, foreach, dataclasses, and import, export, instanceof, and new.

The links above reference existing implementations or languages that inspired each feature.

I'm looking for contributors interested in compiler design, language design, static analysis, or developer tooling. Any help is welcome.
The compiler will be written in LuaJIT (or in standard Lua).
Even feedback and design suggestions are greatly appreciated.

---

I'm not sure if I chose the right forum. If I didn't choose the right one, I will move this thread to another forum.

Click to expand...

Unfortunately, I might cancel it; no one wants to help or offer ideas.

I am not sure how much benefit this will bring to mod development when the SRB2 Lua API is already straightforward enough for that specific purpose.

Is this supposed to boost performance for instance? Is this intended to be an addon written with Lua or will it be a compiled plugin that can link to the executable? There is currently a feature list and an idea that it could improve mod development, but unless it also comes with a design and an architecture that makes it clear as to how it will work and what the approach will be, it would not be wise to just get started at this moment.