Welcome to Ygen! This repository contains the source code of the ygen project.
Ygen is a toolkit for building modern compilers, using a llvm like api.
You are probably wondering: Why would I choose ygen and not llvm? Here are a few reasons:
[!WARNING] This project is still early in its developement. Bugs and miscompilations are expected.
ONLY USE YGEN FOR TOY COMPILERS
Here is a simple example on how to use Ygen to build an add function:
use std::error::Error;
use Ygen::prelude::*;
pub fn main() -> Result<(), Box<dyn Error>> {
let mut module = Module();
let ty = FnTy(vec![TypeMetadata::i32, TypeMetadata::i32], TypeMetadata::i32);
let func = module.add(
"add", &ty
);
func.extrn();
func.addBlock("entry");
let val = func.BuildAdd(ty.arg(0), ty.arg(1));
func.BuildRet( val );
module.verify()?;
// prints out the ir of the module
println!("{}", module.dump());
let triple = Triple::host();
// compiles the module in the host assembly and saves it in the specified path
module.emitToAsmFile(
triple,
&mut initializeAllTargets(triple)?,
Path::new("out.asm")
)?;
// compiles the module to a host object file
module
.emitMachineCode(
triple,
&mut initializeAllTargets(triple)?,
false // is debugging metadata enabled
)?.0.emit(
OpenOptions::new().write(true).truncate(true).create(true).open("out.o")?,
None // if debugging metadata is enabled here is the outputed metadata
)?;
Ok(())
}
When executed this simple program builds an add function and dumps it's ir:
define i32 @add( i32 %0, i32 %1 ) {
entry:
%2 = add i32 %0, %1
ret i32 %2
}
Ygen currently supports following architectures:
x86-64
wasm
[WIP]This project is owned by Cr0a3 and licensed under the Apache2 License