Haxe and Nim support

[murrtor]

Would be nice to be able to transpile to haxe, to have more targets (not as a library, but as a complete fully working program), as haxe is mainly for complete apps, supports way more targets (like gdscript), and will allow to test performance of different compilation implementations (maybe haxe will run faster in rust/cpp/c, maybe temper, etc.).

and would also be nice to transpile to nim (nim is very flexible).

[tjpalmer]

I agree your thoughts on Haxe are interesting. If you want to try working on a backend, here is some documentation on writing one. It's nontrivial, but I suspect haxe wouldn't take much new machinery.

I also like Nim as a language. Maybe easier to track as a separate issue? Or maybe it will matter more once someone picks up the task.

(We have limited bandwidth on the core team, but Shaw is making good progress on the C++ backend at the moment.)

[murrtor]

Maybe easier to track as a separate issue? Or maybe it will matter more once someone picks up the task.

ye i think it will matter when someone picks it up, and for now, i think, for now, as there's no contributors yet, that would implement it, having support for 2 languages as one issue would be better?
like, easier to find this issue, more popularity it will gain, more chances that someone will pick one or both.

[mikesamuel]

nim fits nicely into our OO style type definitions and its arc mode should mesh well with our MM

[tjpalmer]

nim fits nicely into our OO style type definitions and its arc mode should mesh well with our MM

And while we could support arc with our plans, orc is the default. I guess people would be able to choose either for (potential future) temper-built nim libraries.