I maintain Half-Life 2: Sandbox and lclass, and authored some other projects on the web that use Lua.
There's really PUC-Rio Lua 5.1.5/LuaJIT and PUC-Rio Lua latest. If you use LuaJIT, you're building it against a drop-in replacement for Lua 5.1.5, and that's been the case for the longest time.
However, if there's anything the Lua community needs to learn to move on from, it's maybe the language itself depending on how you're using it.
We later built https://github.com/Planimeter/game-engine-2d, but in reality, the programming language is too slow or too limited (this goes for JavaScript, too) in either implementation to do more sophisticated things that engineers should really just use C or C++ for and consider moving wholesale to WebAssembly.
You'll end up training a better future generation of engineers by doing this, too. Lua has been great for us, really instrumental in accomplishing some cool stuff, but we've been eyeballing WebAssembly for the future.
There's really PUC-Rio Lua 5.1.5/LuaJIT and PUC-Rio Lua latest. If you use LuaJIT, you're building it against a drop-in replacement for Lua 5.1.5, and that's been the case for the longest time.
However, if there's anything the Lua community needs to learn to move on from, it's maybe the language itself depending on how you're using it.
We later built https://github.com/Planimeter/game-engine-2d, but in reality, the programming language is too slow or too limited (this goes for JavaScript, too) in either implementation to do more sophisticated things that engineers should really just use C or C++ for and consider moving wholesale to WebAssembly.
You'll end up training a better future generation of engineers by doing this, too. Lua has been great for us, really instrumental in accomplishing some cool stuff, but we've been eyeballing WebAssembly for the future.