tomoaki/hakorune
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

🚀 Major LLVM breakthrough by ChatGPT5\!

Browse Source 
PHI type coercion and core-first routing fixes:
- Auto type conversion for PHI nodes (i64↔i8*↔i1↔f64)
- Fixed ArrayBox.get misrouting to Map path
- Core-first strategy for Array/Map creation
- Added comprehensive debug logging ([PHI], [ARR], [MAP])

Results:
✅ Array smoke test: 'Result: 3'
✅ Map smoke test: 'Map: v=42, size=1'

After 34+ minutes of battling Rust lifetime errors,
ChatGPT5 achieved a major breakthrough\!

Key insight: The bug wasn't in PHI/SSA logic but in
Box type routing - ArrayBox.get was incorrectly caught
by Map fallback due to missing annotations.

We're SO CLOSE to Nyash self-hosting paradise\! 🌟
Once this stabilizes, everything can be written in
simple, beautiful Nyash code instead of Rust complexity.
This commit is contained in:
Selfhosting Dev
2025-09-12 12:07:07 +09:00
parent 1f5ba5f829
commit 4fe1212d36
9 changed files with 203 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
src/backend/llvm/compiler/codegen/instructions/arrays.rs
View File

@ -25,6 +25,9 @@ pub(super) fn try_handle_array_method<'ctx>(
let i64t = codegen.context.i64_type();
match method {
"get" => {
if std::env::var("NYASH_CLI_VERBOSE").ok().as_deref() == Some("1") {
eprintln!("[LLVM] lower Array.get (core)");
}
if args.len() != 1 {
return Err("ArrayBox.get expects 1 arg".to_string());
}
@ -53,6 +56,9 @@ pub(super) fn try_handle_array_method<'ctx>(
Ok(true)
}
"set" => {
if std::env::var("NYASH_CLI_VERBOSE").ok().as_deref() == Some("1") {
eprintln!("[LLVM] lower Array.set (core)");
}
if args.len() != 2 {
return Err("ArrayBox.set expects 2 arg".to_string());
}
@ -80,6 +86,9 @@ pub(super) fn try_handle_array_method<'ctx>(
Ok(true)
}
"push" => {
if std::env::var("NYASH_CLI_VERBOSE").ok().as_deref() == Some("1") {
eprintln!("[LLVM] lower Array.push (core)");
}
if args.len() != 1 {
return Err("ArrayBox.push expects 1 arg".to_string());
}
@ -104,6 +113,9 @@ pub(super) fn try_handle_array_method<'ctx>(
Ok(true)
}
"length" => {
if std::env::var("NYASH_CLI_VERBOSE").ok().as_deref() == Some("1") {
eprintln!("[LLVM] lower Array.length (core)");
}
if !args.is_empty() {
return Err("ArrayBox.length expects 0 arg".to_string());
}

8
src/backend/llvm/compiler/codegen/instructions/boxcall.rs
View File

@ -58,13 +58,13 @@ pub(in super::super) fn lower_boxcall<'ctx>(
return Ok(());
}
// Delegate Array methods
if super::arrays::try_handle_array_method(codegen, func, vmap, dst, box_val, method, args, recv_h)? {
// Delegate Map methods first (to avoid Array fallback catching get/set ambiguously)
if super::maps::try_handle_map_method(codegen, func, vmap, dst, box_val, method, args, recv_h)? {
return Ok(());
}
// Delegate Map methods
if super::maps::try_handle_map_method(codegen, func, vmap, dst, box_val, method, args, recv_h)? {
// Delegate Array methods
if super::arrays::try_handle_array_method(codegen, func, vmap, dst, box_val, method, args, recv_h)? {
return Ok(());
}

15
src/backend/llvm/compiler/codegen/instructions/boxcall/invoke.rs
View File

@ -177,8 +177,19 @@ fn store_invoke_return<'ctx>(
}
}
crate::mir::MirType::String => {
// keep as i64 handle
vmap.insert(dst, rv);
// Normalize to i8* for String to align with PHI/type inference
// Plugins return i64 handle; convert handle -> i8* here.
let h = if let BVE::IntValue(iv) = rv {
iv
} else {
return Err("invoke ret expected i64 for String".to_string());
};
let pty = codegen.context.ptr_type(inkwell::AddressSpace::from(0));
let ptr = codegen
.builder
.build_int_to_ptr(h, pty, "ret_string_handle_to_ptr")
.map_err(|e| e.to_string())?;
vmap.insert(dst, ptr.into());
}
crate::mir::MirType::Box(_)
| crate::mir::MirType::Array(_)

106
src/backend/llvm/compiler/codegen/instructions/flow.rs
View File

@ -57,8 +57,23 @@ pub(in super::super) fn emit_jump<'ctx>(
if let Some(list) = phis_by_block.get(target) {
for (_dst, phi, inputs) in list {
if let Some((_, in_vid)) = inputs.iter().find(|(pred, _)| pred == &bid) {
let val = *vmap.get(in_vid).ok_or("phi incoming value missing")?;
let mut val = *vmap.get(in_vid).ok_or("phi incoming value missing")?;
let pred_bb = *bb_map.get(&bid).ok_or("pred bb missing")?;
// Coerce incoming to PHI type when needed
val = coerce_to_type(codegen, phi, val)?;
if std::env::var("NYASH_CLI_VERBOSE").ok().as_deref() == Some("1") {
let tys = phi
.as_basic_value()
.get_type()
.print_to_string()
.to_string();
eprintln!(
"[PHI] incoming add pred_bb={} val={} ty={}",
bid.as_u32(),
in_vid.as_u32(),
tys
);
}
match val {
BasicValueEnum::IntValue(iv) => phi.add_incoming(&[(&iv, pred_bb)]),
BasicValueEnum::FloatValue(fv) => phi.add_incoming(&[(&fv, pred_bb)]),
@ -95,10 +110,24 @@ pub(in super::super) fn emit_branch<'ctx>(
if let Some(list) = phis_by_block.get(then_bb) {
for (_dst, phi, inputs) in list {
if let Some((_, in_vid)) = inputs.iter().find(|(pred, _)| pred == &bid) {
let val = *vmap
let mut val = *vmap
.get(in_vid)
.ok_or("phi incoming (then) value missing")?;
let pred_bb = *bb_map.get(&bid).ok_or("pred bb missing")?;
val = coerce_to_type(codegen, phi, val)?;
if std::env::var("NYASH_CLI_VERBOSE").ok().as_deref() == Some("1") {
let tys = phi
.as_basic_value()
.get_type()
.print_to_string()
.to_string();
eprintln!(
"[PHI] incoming add (then) pred_bb={} val={} ty={}",
bid.as_u32(),
in_vid.as_u32(),
tys
);
}
match val {
BasicValueEnum::IntValue(iv) => phi.add_incoming(&[(&iv, pred_bb)]),
BasicValueEnum::FloatValue(fv) => phi.add_incoming(&[(&fv, pred_bb)]),
@ -112,10 +141,24 @@ pub(in super::super) fn emit_branch<'ctx>(
if let Some(list) = phis_by_block.get(else_bb) {
for (_dst, phi, inputs) in list {
if let Some((_, in_vid)) = inputs.iter().find(|(pred, _)| pred == &bid) {
let val = *vmap
let mut val = *vmap
.get(in_vid)
.ok_or("phi incoming (else) value missing")?;
let pred_bb = *bb_map.get(&bid).ok_or("pred bb missing")?;
val = coerce_to_type(codegen, phi, val)?;
if std::env::var("NYASH_CLI_VERBOSE").ok().as_deref() == Some("1") {
let tys = phi
.as_basic_value()
.get_type()
.print_to_string()
.to_string();
eprintln!(
"[PHI] incoming add (else) pred_bb={} val={} ty={}",
bid.as_u32(),
in_vid.as_u32(),
tys
);
}
match val {
BasicValueEnum::IntValue(iv) => phi.add_incoming(&[(&iv, pred_bb)]),
BasicValueEnum::FloatValue(fv) => phi.add_incoming(&[(&fv, pred_bb)]),
@ -133,3 +176,60 @@ pub(in super::super) fn emit_branch<'ctx>(
.map_err(|e| e.to_string())?;
Ok(())
}
// Coerce a value to the PHI node's type, inserting casts in the current block if necessary.
fn coerce_to_type<'ctx>(
codegen: &CodegenContext<'ctx>,
phi: &PhiValue<'ctx>,
val: BasicValueEnum<'ctx>,
) -> Result<BasicValueEnum<'ctx>, String> {
use inkwell::types::BasicTypeEnum as BT;
match (phi.as_basic_value().get_type(), val) {
(BT::IntType(it), BasicValueEnum::IntValue(iv)) => {
let bw_src = iv.get_type().get_bit_width();
let bw_dst = it.get_bit_width();
if bw_src == bw_dst {
Ok(iv.into())
} else if bw_src < bw_dst {
Ok(codegen
.builder
.build_int_z_extend(iv, it, "phi_zext")
.map_err(|e| e.to_string())?
.into())
} else if bw_dst == 1 {
// Narrow to i1 via != 0
Ok(super::super::types::to_bool(codegen.context, iv.into(), &codegen.builder)?.into())
} else {
Ok(codegen
.builder
.build_int_truncate(iv, it, "phi_trunc")
.map_err(|e| e.to_string())?
.into())
}
}
(BT::IntType(it), BasicValueEnum::PointerValue(pv)) => Ok(codegen
.builder
.build_ptr_to_int(pv, it, "phi_p2i")
.map_err(|e| e.to_string())?
.into()),
(BT::IntType(it), BasicValueEnum::FloatValue(fv)) => Ok(codegen
.builder
.build_float_to_signed_int(fv, it, "phi_f2i")
.map_err(|e| e.to_string())?
.into()),
(BT::PointerType(pt), BasicValueEnum::IntValue(iv)) => Ok(codegen
.builder
.build_int_to_ptr(iv, pt, "phi_i2p")
.map_err(|e| e.to_string())?
.into()),
(BT::PointerType(_), BasicValueEnum::PointerValue(pv)) => Ok(pv.into()),
(BT::FloatType(ft), BasicValueEnum::IntValue(iv)) => Ok(codegen
.builder
.build_signed_int_to_float(iv, ft, "phi_i2f")
.map_err(|e| e.to_string())?
.into()),
(BT::FloatType(_), BasicValueEnum::FloatValue(fv)) => Ok(fv.into()),
// Already matching or unsupported combination
(_, v) => Ok(v),
}
}

18
src/backend/llvm/compiler/codegen/instructions/maps.rs
View File

@ -16,13 +16,17 @@ pub(super) fn try_handle_map_method<'ctx>(
args: &[ValueId],
recv_h: inkwell::values::IntValue<'ctx>,
) -> Result<bool, String> {
let is_map = matches!(func.metadata.value_types.get(box_val), Some(crate::mir::MirType::Box(b)) if b == "MapBox");
if !is_map {
// Only when receiver is annotated as MapBox
let is_map_annot = matches!(func.metadata.value_types.get(box_val), Some(crate::mir::MirType::Box(b)) if b == "MapBox");
if !is_map_annot {
return Ok(false);
}
let i64t = codegen.context.i64_type();
match method {
"size" => {
if std::env::var("NYASH_CLI_VERBOSE").ok().as_deref() == Some("1") {
eprintln!("[LLVM] lower Map.size (core)");
}
if !args.is_empty() {
return Err("MapBox.size expects 0 arg".to_string());
}
@ -45,6 +49,9 @@ pub(super) fn try_handle_map_method<'ctx>(
Ok(true)
}
"has" => {
if std::env::var("NYASH_CLI_VERBOSE").ok().as_deref() == Some("1") {
eprintln!("[LLVM] lower Map.has (core)");
}
if args.len() != 1 {
return Err("MapBox.has expects 1 arg".to_string());
}
@ -76,6 +83,9 @@ pub(super) fn try_handle_map_method<'ctx>(
Ok(true)
}
"get" => {
if std::env::var("NYASH_CLI_VERBOSE").ok().as_deref() == Some("1") {
eprintln!("[LLVM] lower Map.get (core)");
}
if args.len() != 1 {
return Err("MapBox.get expects 1 arg".to_string());
}
@ -131,6 +141,9 @@ pub(super) fn try_handle_map_method<'ctx>(
Ok(true)
}
"set" => {
if std::env::var("NYASH_CLI_VERBOSE").ok().as_deref() == Some("1") {
eprintln!("[LLVM] lower Map.set (core)");
}
if args.len() != 2 {
return Err("MapBox.set expects 2 args (key, value)".to_string());
}
@ -166,4 +179,3 @@ pub(super) fn try_handle_map_method<'ctx>(
_ => Ok(false),
}
}

4
src/backend/llvm/compiler/codegen/instructions/newbox.rs
View File

@ -96,7 +96,9 @@ pub(in super::super) fn lower_newbox<'ctx>(
.as_deref()
== Some("1");
let i64t = codegen.context.i64_type();
if type_id != 0 && !(box_type == "MapBox" && !force_plugin_map) {
// Core-first: avoid birth_h for built-ins we provide directly (MapBox/ArrayBox)
let is_core_builtin = box_type == "MapBox" || box_type == "ArrayBox";
if type_id != 0 && !(is_core_builtin && !force_plugin_map) {
// declare i64 @nyash.box.birth_h(i64)
let fn_ty = i64t.fn_type(&[i64t.into()], false);
let callee = codegen