Refactor: Pass global names to Binder

The `Binder` struct now accepts a `Rc<RefCell<HashMap<Symbol,
(GlobalIdx, Identity)>>>` for global names. This allows the binder to
directly access and manage global symbols during the binding process,
simplifying the logic and improving efficiency.

Additionally, `CompilerScope` now implements `Default`, and a type alias
`BindingResult` has been introduced for clarity. The `bind_root`
function signature has been updated to accept the `globals` argument.
This commit is contained in:
Michael Schimmel
2026-03-12 16:21:20 +01:00
parent a220815bd6
commit dcb7685d29
4 changed files with 86 additions and 28 deletions
+56 -14
View File
@@ -21,6 +21,12 @@ pub struct CompilerScope {
pub locals: HashMap<Symbol, LocalInfo>, pub locals: HashMap<Symbol, LocalInfo>,
} }
impl Default for CompilerScope {
fn default() -> Self {
Self::new()
}
}
impl CompilerScope { impl CompilerScope {
pub fn new() -> Self { pub fn new() -> Self {
Self { Self {
@@ -108,8 +114,16 @@ pub enum ExprContext {
Statement, Statement,
} }
pub type BindingResult = (
BoundNode,
HashMap<Identity, Vec<Identity>>,
Vec<CompilerScope>,
u32,
);
pub struct Binder { pub struct Binder {
functions: Vec<FunctionCompiler>, functions: Vec<FunctionCompiler>,
globals: Rc<RefCell<HashMap<Symbol, (GlobalIdx, Identity)>>>,
capture_map: HashMap<Identity, std::collections::HashSet<Identity>>, capture_map: HashMap<Identity, std::collections::HashSet<Identity>>,
fixed_scope_idx: i32, fixed_scope_idx: i32,
} }
@@ -119,9 +133,11 @@ impl Binder {
initial_scopes: Vec<CompilerScope>, initial_scopes: Vec<CompilerScope>,
initial_slot_count: u32, initial_slot_count: u32,
fixed_scope_idx: i32, fixed_scope_idx: i32,
globals: Rc<RefCell<HashMap<Symbol, (GlobalIdx, Identity)>>>,
) -> Self { ) -> Self {
let mut binder = Self { let mut binder = Self {
functions: Vec::new(), functions: Vec::new(),
globals,
capture_map: HashMap::new(), capture_map: HashMap::new(),
fixed_scope_idx, fixed_scope_idx,
}; };
@@ -140,10 +156,11 @@ impl Binder {
initial_scopes: Vec<CompilerScope>, initial_scopes: Vec<CompilerScope>,
initial_slot_count: u32, initial_slot_count: u32,
fixed_scope_idx: i32, fixed_scope_idx: i32,
globals: Rc<RefCell<HashMap<Symbol, (GlobalIdx, Identity)>>>,
node: &Node<UntypedKind>, node: &Node<UntypedKind>,
diagnostics: &mut Diagnostics, diagnostics: &mut Diagnostics,
) -> Result<(BoundNode, HashMap<Identity, Vec<Identity>>, Vec<CompilerScope>, u32), String> { ) -> Result<BindingResult, String> {
let mut binder = Self::new(initial_scopes, initial_slot_count, fixed_scope_idx); let mut binder = Self::new(initial_scopes, initial_slot_count, fixed_scope_idx, globals);
let bound = binder.bind(node, ExprContext::Expression, diagnostics); let bound = binder.bind(node, ExprContext::Expression, diagnostics);
let final_captures = binder let final_captures = binder
@@ -531,14 +548,25 @@ impl Binder {
// 2. Try enclosing scopes (capture chain) // 2. Try enclosing scopes (capture chain)
for i in (0..current_fn_idx).rev() { for i in (0..current_fn_idx).rev() {
if let Some((info, _)) = self.functions[i].resolve_local(sym) { if let Some((info, scope_idx)) = self.functions[i].resolve_local(sym) {
// If the resolved address is already Global, we don't need to capture it as an upvalue // If it's in the root script and within a frozen scope, translate to Global
let is_frozen_root = i == 0 && (scope_idx as i32) <= self.fixed_scope_idx;
if let Address::Global(_) = info.addr { if let Address::Global(_) = info.addr {
return Some(info.addr); return Some(info.addr);
} }
let mut addr = info.addr; let mut addr = info.addr;
if is_frozen_root
&& let Address::Local(slot) = addr {
addr = Address::Global(GlobalIdx(slot.0));
}
if let Address::Global(_) = addr {
return Some(addr);
}
// Record the capture for each lambda level in between // Record the capture for each lambda level in between
for k in (i + 1)..=current_fn_idx { for k in (i + 1)..=current_fn_idx {
let lambda_id = self.functions[k].identity.clone(); let lambda_id = self.functions[k].identity.clone();
@@ -552,7 +580,13 @@ impl Binder {
} }
} }
// 3. Global Fallback (search in root level with context removed) // 3. Global Fallback
let globals = self.globals.borrow();
if let Some((idx, _)) = globals.get(sym) {
return Some(Address::Global(*idx));
}
// 4. Global Fallback (search in root level with context removed)
if sym.context.is_some() { if sym.context.is_some() {
let fallback_sym = Symbol { let fallback_sym = Symbol {
name: sym.name.clone(), name: sym.name.clone(),
@@ -560,12 +594,15 @@ impl Binder {
}; };
// Search again in all functions, primarily we care about Root Scopes // Search again in all functions, primarily we care about Root Scopes
for i in (0..=current_fn_idx).rev() { for i in (0..=current_fn_idx).rev() {
if let Some((info, _)) = self.functions[i].resolve_local(&fallback_sym) { if let Some((info, _)) = self.functions[i].resolve_local(&fallback_sym)
if let Address::Global(_) = info.addr { && let Address::Global(_) = info.addr
return Some(info.addr); {
} return Some(info.addr);
} }
} }
if let Some((idx, _)) = globals.get(&fallback_sym) {
return Some(Address::Global(*idx));
}
} }
diag.push_error( diag.push_error(
@@ -682,8 +719,9 @@ mod tests {
let mut parser = Parser::new(source); let mut parser = Parser::new(source);
let untyped = parser.parse_expression(); let untyped = parser.parse_expression();
let globals = Rc::new(RefCell::new(HashMap::new()));
let mut diagnostics = Diagnostics::new(); let mut diagnostics = Diagnostics::new();
let (bound, captures, _scopes, _slots) = Binder::bind_root(vec![], 0, 0, &untyped, &mut diagnostics).unwrap(); let (bound, captures, _scopes, _slots) = Binder::bind_root(vec![], 0, 0, globals, &untyped, &mut diagnostics).unwrap();
if let BoundKind::Lambda { body, .. } = &bound.kind { if let BoundKind::Lambda { body, .. } = &bound.kind {
if let BoundKind::Block { exprs } = &body.kind { if let BoundKind::Block { exprs } = &body.kind {
@@ -711,8 +749,9 @@ mod tests {
let mut parser = Parser::new(source); let mut parser = Parser::new(source);
let untyped = parser.parse_expression(); let untyped = parser.parse_expression();
let globals = Rc::new(RefCell::new(HashMap::new()));
let mut diagnostics = Diagnostics::new(); let mut diagnostics = Diagnostics::new();
let (bound, captures, _scopes, _slots) = Binder::bind_root(vec![], 0, 0, &untyped, &mut diagnostics).unwrap(); let (bound, captures, _scopes, _slots) = Binder::bind_root(vec![], 0, 0, globals, &untyped, &mut diagnostics).unwrap();
if let BoundKind::Lambda { body, .. } = &bound.kind { if let BoundKind::Lambda { body, .. } = &bound.kind {
if let BoundKind::Block { exprs } = &body.kind { if let BoundKind::Block { exprs } = &body.kind {
@@ -740,8 +779,9 @@ mod tests {
let mut parser = Parser::new(source); let mut parser = Parser::new(source);
let untyped = parser.parse_expression(); let untyped = parser.parse_expression();
let globals = Rc::new(RefCell::new(HashMap::new()));
let mut diagnostics = Diagnostics::new(); let mut diagnostics = Diagnostics::new();
let _ = Binder::bind_root(vec![], 0, 0, &untyped, &mut diagnostics); let _ = Binder::bind_root(vec![], 0, 0, globals, &untyped, &mut diagnostics);
assert!(diagnostics.has_errors()); assert!(diagnostics.has_errors());
assert!(diagnostics.items.iter().any(|i| i.message.contains("already defined"))); assert!(diagnostics.items.iter().any(|i| i.message.contains("already defined")));
@@ -749,16 +789,18 @@ mod tests {
#[test] #[test]
fn test_repro_global_redefinition() { fn test_repro_global_redefinition() {
let globals = Rc::new(RefCell::new(HashMap::new()));
let source1 = "(def x 1) 1"; let source1 = "(def x 1) 1";
let untyped1 = Parser::new(source1).parse_expression(); let untyped1 = Parser::new(source1).parse_expression();
let mut diagnostics = Diagnostics::new(); let mut diagnostics = Diagnostics::new();
let (_, _, scopes1, slots1) = Binder::bind_root(vec![], 0, -1, &untyped1, &mut diagnostics).unwrap(); let (_, _, scopes1, slots1) = Binder::bind_root(vec![], 0, -1, globals.clone(), &untyped1, &mut diagnostics).unwrap();
let source2 = "(def x 2) 2"; let source2 = "(def x 2) 2";
let untyped2 = Parser::new(source2).parse_expression(); let untyped2 = Parser::new(source2).parse_expression();
let mut diagnostics2 = Diagnostics::new(); let mut diagnostics2 = Diagnostics::new();
// Here we simulate frozen scope by passing fixed_scope_idx = 0 // Here we simulate frozen scope by passing fixed_scope_idx = 0
let _ = Binder::bind_root(scopes1, slots1, 0, &untyped2, &mut diagnostics2); let _ = Binder::bind_root(scopes1, slots1, 0, globals.clone(), &untyped2, &mut diagnostics2);
assert!(diagnostics2.has_errors()); assert!(diagnostics2.has_errors());
assert!(diagnostics2.items.iter().any(|i| i.message.contains("frozen/immutable"))); assert!(diagnostics2.items.iter().any(|i| i.message.contains("frozen/immutable")));
+3 -3
View File
@@ -783,7 +783,7 @@ mod tests {
let mut diag = crate::ast::diagnostics::Diagnostics::new(); let mut diag = crate::ast::diagnostics::Diagnostics::new();
// fixed_scope_idx = 0 means scope 0 is frozen (Global) // fixed_scope_idx = 0 means scope 0 is frozen (Global)
let result = Binder::bind_root(initial_scopes, 1, 0, &expanded, &mut diag); let result = Binder::bind_root(initial_scopes, 1, 0, Rc::new(RefCell::new(HashMap::new())), &expanded, &mut diag);
assert!( assert!(
result.is_ok(), result.is_ok(),
"Should find global '*' Error: {:?}", "Should find global '*' Error: {:?}",
@@ -807,7 +807,7 @@ mod tests {
let expanded = expander.expand(untyped).unwrap(); let expanded = expander.expand(untyped).unwrap();
let mut diag = crate::ast::diagnostics::Diagnostics::new(); let mut diag = crate::ast::diagnostics::Diagnostics::new();
let result = Binder::bind_root(vec![], 0, 0, &expanded, &mut diag); let result = Binder::bind_root(vec![], 0, 0, Rc::new(RefCell::new(HashMap::new())), &expanded, &mut diag);
assert!(result.is_ok(), "Hygiene failed: {:?}", result.err()); assert!(result.is_ok(), "Hygiene failed: {:?}", result.err());
} }
@@ -827,7 +827,7 @@ mod tests {
let expanded = expander.expand(untyped).unwrap(); let expanded = expander.expand(untyped).unwrap();
let mut diag = crate::ast::diagnostics::Diagnostics::new(); let mut diag = crate::ast::diagnostics::Diagnostics::new();
let result = Binder::bind_root(vec![], 0, 0, &expanded, &mut diag); let result = Binder::bind_root(vec![], 0, 0, Rc::new(RefCell::new(HashMap::new())), &expanded, &mut diag);
assert!( assert!(
result.is_ok(), result.is_ok(),
"Explicit Hygiene with Backticks failed: {:?}", "Explicit Hygiene with Backticks failed: {:?}",
+1 -1
View File
@@ -717,7 +717,7 @@ mod tests {
assert!(result.is_err()); assert!(result.is_err());
assert_eq!( assert_eq!(
result.unwrap_err(), result.unwrap_err(),
"Statement 'def' cannot be used as an expression.\nCannot define 'x': Scope 0 is frozen/immutable.\nCannot destructure type int as a tuple/vector" "Statement 'def' cannot be used as an expression.\nCannot destructure type int as a tuple/vector"
); );
} }
+26 -10
View File
@@ -112,6 +112,7 @@ impl FunctionRegistry for EnvFunctionRegistry {
} }
struct RuntimeMacroEvaluator { struct RuntimeMacroEvaluator {
global_names: Rc<RefCell<HashMap<Symbol, (GlobalIdx, Identity)>>>,
root_scopes: Rc<RefCell<Vec<crate::ast::compiler::binder::CompilerScope>>>, root_scopes: Rc<RefCell<Vec<crate::ast::compiler::binder::CompilerScope>>>,
root_slot_count: Rc<RefCell<u32>>, root_slot_count: Rc<RefCell<u32>>,
global_types: Rc<RefCell<HashMap<GlobalIdx, StaticType>>>, global_types: Rc<RefCell<HashMap<GlobalIdx, StaticType>>>,
@@ -135,7 +136,7 @@ impl MacroEvaluator for RuntimeMacroEvaluator {
let initial_scopes = self.root_scopes.borrow().clone(); let initial_scopes = self.root_scopes.borrow().clone();
let initial_slot_count = *self.root_slot_count.borrow(); let initial_slot_count = *self.root_slot_count.borrow();
let (bound_ast, captures, _, _) = Binder::bind_root(initial_scopes, initial_slot_count, self.fixed_scope_idx, node, &mut diag)?; let (bound_ast, captures, _, _) = Binder::bind_root(initial_scopes, initial_slot_count, self.fixed_scope_idx, self.global_names.clone(), node, &mut diag)?;
let bound_ast = crate::ast::compiler::captures::CapturePass::apply(bound_ast, &captures); let bound_ast = crate::ast::compiler::captures::CapturePass::apply(bound_ast, &captures);
let checker = TypeChecker::new(self.global_types.clone()); let checker = TypeChecker::new(self.global_types.clone());
let typed_ast = checker.check(&bound_ast, &[], &mut diag); let typed_ast = checker.check(&bound_ast, &[], &mut diag);
@@ -186,6 +187,8 @@ impl Environment {
let mut env = env; let mut env = env;
env.fixed_scope_idx = 0; env.fixed_scope_idx = 0;
// Push the first mutable user scope (Level 1)
env.root_scopes.borrow_mut().push(crate::ast::compiler::binder::CompilerScope::new());
// Automatically add standard search paths (CWD and CWD/rtl) // Automatically add standard search paths (CWD and CWD/rtl)
if let Ok(cwd) = std::env::current_dir() { if let Ok(cwd) = std::env::current_dir() {
@@ -223,6 +226,7 @@ impl Environment {
fn get_expander(&self) -> MacroExpander<RuntimeMacroEvaluator> { fn get_expander(&self) -> MacroExpander<RuntimeMacroEvaluator> {
let evaluator = RuntimeMacroEvaluator { let evaluator = RuntimeMacroEvaluator {
global_names: self.global_names.clone(),
root_scopes: self.root_scopes.clone(), root_scopes: self.root_scopes.clone(),
root_slot_count: self.root_slot_count.clone(), root_slot_count: self.root_slot_count.clone(),
global_types: self.global_types.clone(), global_types: self.global_types.clone(),
@@ -398,10 +402,23 @@ impl Environment {
match &node.kind { match &node.kind {
UntypedKind::Def { target, .. } => { UntypedKind::Def { target, .. } => {
if let UntypedKind::Identifier(sym) = &target.kind { if let UntypedKind::Identifier(sym) = &target.kind {
let mut names = self.global_names.borrow_mut(); let mut root_scopes = self.root_scopes.borrow_mut();
if !names.contains_key(sym) { let last_idx = root_scopes.len() - 1;
let idx = GlobalIdx(names.len() as u32); let current_scope = &mut root_scopes[last_idx];
names.insert(sym.clone(), (idx, node.identity.clone()));
if !current_scope.locals.contains_key(sym) {
let mut slot_count = self.root_slot_count.borrow_mut();
let slot = crate::ast::compiler::bound_nodes::LocalSlot(*slot_count);
current_scope.locals.insert(
sym.clone(),
crate::ast::compiler::binder::LocalInfo {
addr: Address::Local(slot),
identity: node.identity.clone(),
_ty: StaticType::Any,
purity: Purity::Impure,
},
);
*slot_count += 1;
} }
} }
} }
@@ -442,7 +459,7 @@ impl Environment {
let initial_slot_count = *self.root_slot_count.borrow(); let initial_slot_count = *self.root_slot_count.borrow();
let (bound_ast, captures, final_scopes, final_slot_count) = let (bound_ast, captures, final_scopes, final_slot_count) =
match Binder::bind_root(initial_scopes, initial_slot_count, self.fixed_scope_idx, &expanded_ast, diagnostics) { match Binder::bind_root(initial_scopes, initial_slot_count, self.fixed_scope_idx, self.global_names.clone(), &expanded_ast, diagnostics) {
Ok(res) => res, Ok(res) => res,
Err(e) => { Err(e) => {
diagnostics.push_error(e, None); diagnostics.push_error(e, None);
@@ -459,10 +476,9 @@ impl Environment {
// Pre-allocate global slots to prevent out-of-bounds during specialization/optimization // Pre-allocate global slots to prevent out-of-bounds during specialization/optimization
{ {
let mut values = self.global_values.borrow_mut(); let mut values = self.global_values.borrow_mut();
let names = self.global_names.borrow(); let count = *self.root_slot_count.borrow();
let max_idx = names.values().map(|(idx, _)| idx.0).max().unwrap_or(0); if (count as usize) > values.len() {
if (max_idx as usize) >= values.len() { values.resize(count as usize, Value::Void);
values.resize((max_idx + 1) as usize, Value::Void);
} }
} }