Files
RustAst/src/ast/vm.rs
T
Michael Schimmel 98deb8f3fe Refactor VM evaluation to use macro
The `eval` and `eval_observed` methods in the `VM` struct were very
similar, with the primary difference being the call to the `VMObserver`
trait. This commit refactors these methods into a single macro,
`dispatch_eval!`, which reduces code duplication and improves
maintainability.

The `Value::TailCallRequest` variant was also updated to use `Box` for
its contents, which helps keep the `Value` enum size consistent.

Finally, the benchmarks in the `examples` directory have been updated to
reflect minor performance changes resulting from these code
modifications.
2026-02-18 02:01:27 +01:00

547 lines
20 KiB
Rust

use std::rc::Rc;
use std::cell::RefCell;
use std::collections::HashMap;
use std::any::Any;
use crate::ast::compiler::bound_nodes::{Address, BoundKind};
use crate::ast::nodes::Node;
use crate::ast::types::{Value, Object, StaticType};
#[derive(Debug, Clone)]
pub struct Closure {
pub function_node: Rc<Node<BoundKind, StaticType>>,
pub upvalues: Vec<Rc<RefCell<Value>>>,
}
impl Object for Closure {
fn type_name(&self) -> &'static str {
"closure"
}
fn as_any(&self) -> &dyn Any {
self
}
}
#[derive(Debug)]
struct CallFrame {
stack_base: usize,
closure: Option<Rc<Closure>>,
}
/// Hook for observing VM execution (zero-cost when O::ACTIVE is false)
pub trait VMObserver {
const ACTIVE: bool = false;
fn before_eval(&mut self, _vm: &VM, _node: &Node<BoundKind, StaticType>) {}
fn after_eval(&mut self, _vm: &VM, _node: &Node<BoundKind, StaticType>, _res: &Result<Value, String>) {}
}
/// Default observer that does nothing and should be optimized away
pub struct NoOpObserver;
impl VMObserver for NoOpObserver {}
/// Observer that logs the execution tree and scope state
pub struct TracingObserver {
pub logs: Vec<String>,
indent: usize,
}
impl TracingObserver {
pub fn new() -> Self {
Self { logs: Vec::new(), indent: 0 }
}
fn pad(&self) -> String {
"| ".repeat(self.indent)
}
}
impl Default for TracingObserver {
fn default() -> Self {
Self::new()
}
}
impl VMObserver for TracingObserver {
const ACTIVE: bool = true;
fn before_eval(&mut self, _vm: &VM, node: &Node<BoundKind, StaticType>) {
let pad = self.pad();
self.logs.push(format!("{}{} [{}]: {{", pad, node.kind.display_name(), node.ty));
self.indent += 1;
}
fn after_eval(&mut self, vm: &VM, node: &Node<BoundKind, StaticType>, res: &Result<Value, String>) {
self.indent = self.indent.saturating_sub(1);
let pad = self.pad();
// Show scope state on certain nodes (like Delphi ShowScope)
match &node.kind {
BoundKind::DefLocal { .. } | BoundKind::Set { .. } | BoundKind::DefGlobal { .. } => {
let s_pad = format!("{}| ", pad);
self.logs.push(format!("{}--- Scope Status ---", s_pad));
self.logs.push(format!("{}Stack (top 5): {:?}", s_pad, vm.stack.iter().rev().take(5).collect::<Vec<_>>()));
},
_ => {}
}
let res_str = match res {
Ok(v) => format!("{}", v),
Err(e) => format!("ERROR: {}", e),
};
self.logs.push(format!("{}}} -> {}", pad, res_str));
}
}
pub struct VM {
stack: Vec<Value>,
globals: Rc<RefCell<Vec<Value>>>,
frames: Vec<CallFrame>,
}
macro_rules! dispatch_eval {
($self:ident, $node:ident, $eval_method:ident $(, $observer:ident)?) => {
match &$node.kind {
BoundKind::Nop => Ok(Value::Void),
BoundKind::Constant(v) => Ok(v.clone()),
BoundKind::DefGlobal { global_index, value } => {
let val = $self.$eval_method($($observer,)? value)?;
let idx = *global_index as usize;
let mut globals = $self.globals.borrow_mut();
if idx >= globals.len() {
globals.resize(idx + 1, Value::Void);
}
globals[idx] = val.clone();
Ok(val)
},
BoundKind::Get(addr) => $self.get_value(*addr),
BoundKind::Set { addr, value } => {
let val = $self.$eval_method($($observer,)? value)?;
$self.set_value(*addr, val.clone())?;
Ok(val)
},
BoundKind::DefLocal { slot, value, captured_by } => {
let val = $self.$eval_method($($observer,)? value)?;
let final_val = if !captured_by.is_empty() {
Value::Cell(Rc::new(RefCell::new(val)))
} else {
val
};
let frame = $self.frames.last().ok_or("No call frame")?;
let abs_index = frame.stack_base + (*slot as usize);
if abs_index == $self.stack.len() {
$self.stack.push(final_val.clone());
} else if abs_index < $self.stack.len() {
$self.stack[abs_index] = final_val.clone();
} else {
return Err(format!("Stack gap at local {}", slot));
}
Ok(final_val)
},
BoundKind::If { cond, then_br, else_br } => {
let c = $self.$eval_method($($observer,)? cond)?;
if c.is_truthy() {
$self.$eval_method($($observer,)? then_br)
} else if let Some(e) = else_br {
$self.$eval_method($($observer,)? e)
} else {
Ok(Value::Void)
}
},
BoundKind::Block { exprs } => {
let mut last = Value::Void;
for e in exprs {
last = $self.$eval_method($($observer,)? e)?;
}
Ok(last)
},
BoundKind::Lambda { param_count: _, upvalues, body } => {
let mut captured = Vec::with_capacity(upvalues.len());
for addr in upvalues {
captured.push($self.capture_upvalue(*addr)?);
}
let closure = Closure {
function_node: body.clone(),
upvalues: captured,
};
Ok(Value::Object(Rc::new(closure)))
},
BoundKind::TailCall { callee, args } => {
let func_val = $self.$eval_method($($observer,)? callee)?;
let mut arg_vals = Vec::with_capacity(args.len());
for arg in args {
arg_vals.push($self.$eval_method($($observer,)? arg)?);
}
match func_val {
Value::Object(obj) => Ok(Value::TailCallRequest(Box::new((obj, arg_vals)))),
Value::Function(f) => Ok(f(arg_vals)),
_ => Err(format!("Tail call target is not a function: {}", func_val)),
}
},
BoundKind::Call { callee, args } => {
let mut func_val = $self.$eval_method($($observer,)? callee)?;
let mut arg_vals = Vec::with_capacity(args.len());
for arg in args {
arg_vals.push($self.$eval_method($($observer,)? arg)?);
}
loop {
match func_val {
Value::Function(f) => return Ok(f(arg_vals)),
Value::Object(obj) => {
if let Some(closure) = obj.as_any().downcast_ref::<Closure>() {
let old_stack_top = $self.stack.len();
let closure_rc = Rc::new(closure.clone());
$self.stack.extend(arg_vals);
$self.frames.push(CallFrame {
stack_base: old_stack_top,
closure: Some(closure_rc.clone()),
});
let result = $self.$eval_method($($observer,)? &closure.function_node);
$self.frames.pop();
$self.stack.truncate(old_stack_top);
match result {
Ok(Value::TailCallRequest(payload)) => {
let (next_obj, next_args) = *payload;
func_val = Value::Object(next_obj);
arg_vals = next_args;
continue;
},
Ok(val) => return Ok(val),
Err(e) => return Err(e),
}
} else {
return Err(format!("Object is not a closure: {}", obj.type_name()));
}
},
_ => return Err(format!("Attempt to call non-function: {}", func_val)),
}
}
},
BoundKind::Tuple { elements } => {
let mut vals = Vec::with_capacity(elements.len());
for e in elements {
vals.push($self.$eval_method($($observer,)? e)?);
}
Ok(Value::List(Rc::new(vals)))
},
BoundKind::Map { entries } => {
let mut map = HashMap::new();
for (k, v) in entries {
let key = $self.$eval_method($($observer,)? k)?;
let val = $self.$eval_method($($observer,)? v)?;
if let Value::Keyword(kw) = key {
map.insert(kw, val);
} else {
return Err(format!("Map key must be keyword, got {}", key));
}
}
Ok(Value::Record(Rc::new(map)))
}
}
};
}
impl VM {
pub fn new(globals: Rc<RefCell<Vec<Value>>>) -> Self {
Self {
stack: Vec::new(),
globals,
frames: Vec::new(),
}
}
pub fn run(&mut self, root: &Node<BoundKind, StaticType>) -> Result<Value, String> {
self.stack.clear();
self.frames.clear();
self.frames.push(CallFrame {
stack_base: 0,
closure: None,
});
let result = self.eval(root);
self.frames.pop();
result
}
pub fn run_with_observer<O: VMObserver>(&mut self, observer: &mut O, root: &Node<BoundKind, StaticType>) -> Result<Value, String> {
self.stack.clear();
self.frames.clear();
self.frames.push(CallFrame {
stack_base: 0,
closure: None,
});
let result = self.eval_observed(observer, root);
self.frames.pop();
result
}
/// The pure, original, zero-cost hot path.
#[inline(always)]
fn eval(&mut self, node: &Node<BoundKind, StaticType>) -> Result<Value, String> {
dispatch_eval!(self, node, eval)
}
/// The observed path for debugging.
fn eval_observed<O: VMObserver>(&mut self, observer: &mut O, node: &Node<BoundKind, StaticType>) -> Result<Value, String> {
observer.before_eval(self, node);
let result = dispatch_eval!(self, node, eval_observed, observer);
observer.after_eval(self, node, &result);
result
}
fn capture_upvalue(&mut self, addr: Address) -> Result<Rc<RefCell<Value>>, String> {
match addr {
Address::Local(idx) => {
let frame = self.frames.last().ok_or("No call frame")?;
let abs_index = frame.stack_base + (idx as usize);
if abs_index < self.stack.len() {
// Check if already boxed
if let Value::Cell(cell) = &self.stack[abs_index] {
Ok(cell.clone())
} else {
// Box it
let val = self.stack[abs_index].clone();
let cell = Rc::new(RefCell::new(val));
self.stack[abs_index] = Value::Cell(cell.clone());
Ok(cell)
}
} else {
Err(format!("Stack underflow capture local {}", idx))
}
},
Address::Upvalue(idx) => {
let frame = self.frames.last().ok_or("No call frame")?;
if let Some(closure) = &frame.closure {
let idx = idx as usize;
if idx < closure.upvalues.len() {
Ok(closure.upvalues[idx].clone())
} else {
Err(format!("Upvalue access out of bounds capture {}", idx))
}
} else {
Err("Current frame has no closure".to_string())
}
},
Address::Global(_) => Err("Cannot capture global directly".to_string()),
}
}
fn get_value(&self, addr: Address) -> Result<Value, String> {
match addr {
Address::Local(idx) => {
let frame = self.frames.last().ok_or("No call frame")?;
let abs_index = frame.stack_base + (idx as usize);
if abs_index < self.stack.len() {
match &self.stack[abs_index] {
Value::Cell(cell) => Ok(cell.borrow().clone()),
val => Ok(val.clone()),
}
} else {
Err(format!("Stack underflow access local {}", idx))
}
},
Address::Global(idx) => {
let idx = idx as usize;
let globals = self.globals.borrow();
if idx < globals.len() {
Ok(globals[idx].clone())
} else {
Err(format!("Global access out of bounds {}", idx))
}
},
Address::Upvalue(idx) => {
let frame = self.frames.last().ok_or("No call frame")?;
if let Some(closure) = &frame.closure {
let idx = idx as usize;
if idx < closure.upvalues.len() {
Ok(closure.upvalues[idx].borrow().clone())
} else {
Err(format!("Upvalue access out of bounds {}", idx))
}
} else {
Err("Current frame has no closure (cannot access upvalues)".to_string())
}
}
}
}
fn set_value(&mut self, addr: Address, value: Value) -> Result<(), String> {
match addr {
Address::Local(idx) => {
let frame = self.frames.last().ok_or("No call frame")?;
let abs_index = frame.stack_base + (idx as usize);
if abs_index < self.stack.len() {
if let Value::Cell(cell) = &self.stack[abs_index] {
*cell.borrow_mut() = value;
} else {
self.stack[abs_index] = value;
}
} else if abs_index == self.stack.len() {
self.stack.push(value);
} else {
return Err(format!("Stack gap write local {}", idx));
}
Ok(())
},
Address::Global(idx) => {
let idx = idx as usize;
let mut globals = self.globals.borrow_mut();
if idx >= globals.len() {
globals.resize(idx + 1, Value::Void);
}
globals[idx] = value;
Ok(())
},
Address::Upvalue(idx) => {
let frame = self.frames.last().ok_or("No call frame")?;
if let Some(closure) = &frame.closure {
let idx = idx as usize;
if idx < closure.upvalues.len() {
*closure.upvalues[idx].borrow_mut() = value;
Ok(())
} else {
Err(format!("Upvalue assignment out of bounds {}", idx))
}
} else {
Err("Current frame has no closure".to_string())
}
},
}
}
}
#[cfg(test)]
mod tests {
use super::*;
use crate::ast::types::{SourceLocation, NodeIdentity};
fn make_dummy_identity() -> Rc<NodeIdentity> {
Rc::new(NodeIdentity {
location: SourceLocation { line: 0, col: 0 },
})
}
#[test]
fn test_capture_boxing_modification() {
// Goal:
// var x = 10; (Local 0)
// var f = lambda { x = 20; }; (Local 1)
// f();
// return x; -> Should be 20
let id = make_dummy_identity();
// 1. Define Lambda Body: x = 20 (Upvalue 0)
let lambda_body = Node {
identity: id.clone(),
ty: StaticType::Void,
kind: BoundKind::Set {
addr: Address::Upvalue(0),
value: Box::new(Node {
identity: id.clone(),
ty: StaticType::Int,
kind: BoundKind::Constant(Value::Int(20)),
}),
},
};
// 2. Define Main Script
let root = Node {
identity: id.clone(),
ty: StaticType::Int,
kind: BoundKind::Block {
exprs: vec![
// x = 10 (Local 0) - simulate definition by assignment to stack gap
Node {
identity: id.clone(),
ty: StaticType::Int,
kind: BoundKind::Set {
addr: Address::Local(0),
value: Box::new(Node {
identity: id.clone(),
ty: StaticType::Int,
kind: BoundKind::Constant(Value::Int(10)),
}),
},
},
// f = lambda capturing x (Local 1)
Node {
identity: id.clone(),
ty: StaticType::Any,
kind: BoundKind::Set {
addr: Address::Local(1),
value: Box::new(Node {
identity: id.clone(),
ty: StaticType::Any,
kind: BoundKind::Lambda {
param_count: 0,
upvalues: vec![Address::Local(0)], // Capture x
body: Rc::new(lambda_body),
},
}),
},
},
// f()
Node {
identity: id.clone(),
ty: StaticType::Void,
kind: BoundKind::Call {
callee: Box::new(Node {
identity: id.clone(),
ty: StaticType::Any,
kind: BoundKind::Get(Address::Local(1)),
}),
args: vec![],
},
},
// return x (Local 0)
Node {
identity: id.clone(),
ty: StaticType::Int,
kind: BoundKind::Get(Address::Local(0)),
},
],
},
};
let globals = Rc::new(RefCell::new(Vec::new()));
let mut vm = VM::new(globals);
let result = vm.run(&root);
match result {
Ok(Value::Int(val)) => assert_eq!(val, 20, "Captured variable should be modified to 20"),
Ok(val) => panic!("Expected Int(20), got {:?}", val),
Err(e) => panic!("VM Error: {}", e),
}
}
}