Aside from being an interesting feature, the discussion is a great example of positively encouraging contributions. There's a lot to admire about Edwin Brady: smart, kind, polite, self-effacing, committed. Given the prevalence of acrimony in online discourse - both mainstream and social media - it's always good to be reminded that there are lots of really decent people out there. And anyone who names a language after a dragon in a children's animation[0] gets my vote!
I wonder how much effort would it take to add a rust code generator. I'm guessing we can't add a C code generator because idris seems to need a garbage collector.
Idris 1 compiles via C, it just has to implement a garbage collector in its runtime system. It should be noted that Idris 1 has much worse performance than Idris 2 compiled to Chez Scheme, it would take a serious effort to do better than the current Chez backend.
The logical target for a high-performance backend would be LLVM instead of rust. After elaboration Idris 2 has a series of intermediate languages, from lambda calculus to a simple register machine. Creating LLVM code from the least abstract of them wouldn't be any harder than creating Rust code.
I was thinking of rust specifically because of GC, I don't know how to write a GC and I don't even want to try. But if I use rust, I won't have to because of rust's zero cost abstractions.
There seems to be a bit of confusion here. Idris doesn't provide the same tools and guarantees as Rust does to manage memory, and instead requires that implementations provide a garbage collector.
If you're producing code for a collected language, you can forward on all of the allocations to that language, and if not, you will need to implement one.
Zero-cost abstractions or not, Idris needs a GC from somewhere.
Reference counting would go a long way though as there is very little mutation in a typical Idris program, so you could probably sew something together in Rust with Rc cells and a simple cycle detector.
But basically, having true closures requires some kind of GC. That was already recognized in Steele's Lambda papers in the 70s. Rust has to make compromises with its closures to avoid that.
I was hoping I could get by without resolving to to GC, utilizing the rust allocation model(RAII with syntactic sugar and compiler helping out basically). If I "have to" write a GC because idris programming model requires is, I don't think there is a lot of value in writing a rust code generator.
Haskell used to be compiled to C. (not sure if they still use this backend, I think main GHC backend compiles to LLVM). "Doesn't have garbage collector" is not a good reason not to be able to compile to C, surely garbage collection can be implemented in C.
Just because it is possible doesn't mean it's easy. If you implement garbage collection in C, you are using your own garbage collector, the C ecosystem doesn't support you in any way. When people make the argument of preferring a compilation target because it has a specific feature, they are not saying "target is not Turing complete, you can't implement this feature". They say "not natively supported and I would rather not implement this myself".
Jan Corazza has some really great write-ups about writing bindings to SDL and writing a video game in Idris using them. The source is available on Github
I wrote a Todo application in Idris that compiles to Javascript (https://github.com/leon-vv/Todo). I used the dependent type system of Idris to embed a small part of SQL in Idris, which means the type system can proof certain things about my queries (for example that fields accessed are actually part of the (joined) tables).
Idris is primarily an academic language, much like Haskell. While I don't doubt that people will come to implement such programs in it at some point, I haven't seen anything yet (though I admittedly don't stick close to Idris at the moment so I could've missed something).
I've been interested in Idris for quite awhile, but not sure how to go about learning it. Are there any particularly good resources? Am I expected to know Haskell or is it sufficient to have a few years of experience with FP in general?
You should be good to go picking up "Type-driven development with Idris". It's a great book not just for learning Idris but also for hammering down the concept of using types and the compiler as a design pillar if you're not fully used to it.
[0] https://en.wikipedia.org/wiki/Ivor_the_Engine