formatted

This commit is contained in:
ychenfo 2021-08-11 15:18:21 +08:00
parent 42a636b4ce
commit 99276c8f31
3 changed files with 224 additions and 200 deletions

View File

@ -1,6 +1,6 @@
use crate::location::Location;
use crate::typecheck::typedef::Type;
use crate::top_level::DefinitionId;
use crate::typecheck::typedef::Type;
use rustpython_parser::ast::Expr;
#[derive(Clone, PartialEq)]
@ -21,5 +21,5 @@ pub trait SymbolResolver {
fn get_symbol_value(&self, str: &str) -> Option<SymbolValue>;
fn get_symbol_location(&self, str: &str) -> Option<Location>;
fn get_module_resolver(&self, module_name: &str) -> Option<&dyn SymbolResolver>; // NOTE: for getting imported modules' symbol resolver?
// handle function call etc.
// handle function call etc.
}

View File

@ -27,7 +27,7 @@ pub enum TopLevelDef {
// ancestor classes, including itself.
ancestors: Vec<DefinitionId>,
// symbol resolver of the module defined the class, none if it is built-in type
resolver: Option<Arc<Mutex<dyn SymbolResolver + Send>>>
resolver: Option<Arc<Mutex<dyn SymbolResolver + Send>>>,
},
Function {
// prefix for symbol, should be unique globally, and not ending with numbers
@ -47,7 +47,7 @@ pub enum TopLevelDef {
/// rigid type variables that would be substituted when the function is instantiated.
instance_to_stmt: HashMap<String, (Stmt<Option<Type>>, usize)>,
// symbol resolver of the module defined the class
resolver: Option<Arc<Mutex<dyn SymbolResolver + Send>>>
resolver: Option<Arc<Mutex<dyn SymbolResolver + Send>>>,
},
Initializer {
class_id: DefinitionId,
@ -83,20 +83,19 @@ pub struct CodeGenContext<'ctx> {
pub loop_bb: Option<(BasicBlock<'ctx>, BasicBlock<'ctx>)>,
}
pub fn name_mangling(mut class_name: String, method_name: &str) -> String {
// need to further extend to more name mangling like instantiations of typevar
class_name.push_str(method_name);
class_name
}
pub struct TopLevelDefInfo {
pub struct TopLevelDefInfo {
// like adding some info on top of the TopLevelDef for later parsing the class bases, method,
// and function sigatures
def: TopLevelDef, // the definition entry
ty: Type, // the entry in the top_level unifier
ast: Option<ast::Stmt<()>>, // the ast submitted by applications, primitives and class methods will have None value here
// resolver: Option<&'a dyn SymbolResolver> // the resolver
def: TopLevelDef, // the definition entry
ty: Type, // the entry in the top_level unifier
ast: Option<ast::Stmt<()>>, // the ast submitted by applications, primitives and class methods will have None value here
// resolver: Option<&'a dyn SymbolResolver> // the resolver
}
pub struct TopLevelComposer {
@ -171,28 +170,35 @@ impl TopLevelComposer {
}
/// already include the definition_id of itself inside the ancestors vector
pub fn make_top_level_class_def(index: usize, resolver: Option<Arc<Mutex<dyn SymbolResolver + Send>>>) -> TopLevelDef {
pub fn make_top_level_class_def(
index: usize,
resolver: Option<Arc<Mutex<dyn SymbolResolver + Send>>>,
) -> TopLevelDef {
TopLevelDef::Class {
object_id: DefinitionId(index),
type_vars: Default::default(),
fields: Default::default(),
methods: Default::default(),
ancestors: vec![DefinitionId(index)],
resolver
resolver,
}
}
pub fn make_top_level_function_def(name: String, ty: Type, resolver: Option<Arc<Mutex<dyn SymbolResolver + Send>>>) -> TopLevelDef {
pub fn make_top_level_function_def(
name: String,
ty: Type,
resolver: Option<Arc<Mutex<dyn SymbolResolver + Send>>>,
) -> TopLevelDef {
TopLevelDef::Function {
name,
signature: ty,
instance_to_symbol: Default::default(),
instance_to_stmt: Default::default(),
resolver
resolver,
}
}
// like to make and return a "primitive" symbol resolver? so that the symbol resolver
// like to make and return a "primitive" symbol resolver? so that the symbol resolver
// can later figure out primitive type definitions when passed a primitive type name
pub fn get_primitives_definition(&self) -> Vec<(String, DefinitionId, Type)> {
vec![
@ -207,13 +213,13 @@ impl TopLevelComposer {
pub fn register_top_level(
&mut self,
ast: ast::Stmt<()>,
resolver: Option<Arc<Mutex<dyn SymbolResolver + Send>>>
resolver: Option<Arc<Mutex<dyn SymbolResolver + Send>>>,
) -> Result<Vec<(String, DefinitionId, Type)>, String> {
match &ast.node {
ast::StmtKind::ClassDef { name, body, .. } => {
let class_name = name.to_string();
let class_def_id = self.definition_list.len();
// add the class to the unifier
let ty = self.unifier.add_ty(TypeEnum::TObj {
obj_id: DefinitionId(class_def_id),
@ -221,60 +227,58 @@ impl TopLevelComposer {
params: Default::default(),
});
let mut ret_vector: Vec<(String, DefinitionId, Type)> = vec![(class_name.clone(), DefinitionId(class_def_id), ty)];
let mut ret_vector: Vec<(String, DefinitionId, Type)> =
vec![(class_name.clone(), DefinitionId(class_def_id), ty)];
// parse class def body and register class methods into the def list
// NOTE: module's symbol resolver would not know the name of the class methods,
// thus cannot return their definition_id? so we have to manage it ourselves?
// or do we return the class method list of (method_name, def_id, type) to
// NOTE: module's symbol resolver would not know the name of the class methods,
// thus cannot return their definition_id? so we have to manage it ourselves?
// or do we return the class method list of (method_name, def_id, type) to
// application to be used to build symbol resolver? <- current implementation
// FIXME: better do not return and let symbol resolver to manage the mangled name
for b in body {
if let ast::StmtKind::FunctionDef {name, ..} = &b.node {
if let ast::StmtKind::FunctionDef { name, .. } = &b.node {
let fun_name = name_mangling(class_name.clone(), name);
let def_id = self.definition_list.len();
// add to unifier
let ty = self.unifier.add_ty(TypeEnum::TFunc(crate::typecheck::typedef::FunSignature {
args: Default::default(),
ret: self.primitives.none,
vars: Default::default()
}));
let ty = self.unifier.add_ty(TypeEnum::TFunc(
crate::typecheck::typedef::FunSignature {
args: Default::default(),
ret: self.primitives.none,
vars: Default::default(),
},
));
// add to the definition list
self.definition_list.push(
TopLevelDefInfo {
def: Self::make_top_level_function_def(fun_name.clone(), ty, None), // FIXME:
ty,
ast: None // since it is inside the class def body statments
}
);
self.definition_list.push(TopLevelDefInfo {
def: Self::make_top_level_function_def(fun_name.clone(), ty, None), // FIXME:
ty,
ast: None, // since it is inside the class def body statments
});
ret_vector.push((fun_name, DefinitionId(def_id), ty));
// if it is the contructor, special handling is needed. In the above
// handling, we still add __init__ function to the class method
if name == "__init__" {
self.definition_list.push(
TopLevelDefInfo {
def: TopLevelDef::Initializer {
class_id: DefinitionId(class_def_id)
},
ty: self.primitives.none, // arbitary picked one
ast: None, // it is inside the class def body statments
}
)
self.definition_list.push(TopLevelDefInfo {
def: TopLevelDef::Initializer {
class_id: DefinitionId(class_def_id),
},
ty: self.primitives.none, // arbitary picked one
ast: None, // it is inside the class def body statments
})
// FIXME: should we return this to the symbol resolver?, should be yes
}
} else { } // else do nothing
} else {
} // else do nothing
}
// add to the definition list
self.definition_list.push(
TopLevelDefInfo {
def: Self::make_top_level_class_def(class_def_id, resolver),
ast: Some(ast),
ty,
}
);
self.definition_list.push(TopLevelDefInfo {
def: Self::make_top_level_class_def(class_def_id, resolver),
ast: Some(ast),
ty,
});
Ok(ret_vector)
},
}
ast::StmtKind::FunctionDef { name, .. } => {
let fun_name = name.to_string();
@ -283,18 +287,18 @@ impl TopLevelComposer {
let ty =
self.unifier.add_ty(TypeEnum::TFunc(crate::typecheck::typedef::FunSignature {
args: Default::default(),
ret: self.primitives.none,
vars: Default::default()
}));
ret: self.primitives.none,
vars: Default::default(),
}));
// add to the definition list
self.definition_list.push(TopLevelDefInfo {
def: Self::make_top_level_function_def(
name.into(),
self.primitives.none,
resolver
),
ast: Some(ast),
ty,
def: Self::make_top_level_function_def(
name.into(),
self.primitives.none,
resolver,
),
ast: Some(ast),
ty,
});
Ok(vec![(fun_name, DefinitionId(def_id), ty)])
@ -462,20 +466,25 @@ impl TopLevelComposer {
}
}
pub fn parse_type_var<T>(input: &ast::Expr<T>, resolver: &dyn SymbolResolver) -> Result<Type, String> {
pub fn parse_type_var<T>(
input: &ast::Expr<T>,
resolver: &dyn SymbolResolver,
) -> Result<Type, String> {
match &input.node {
ast::ExprKind::Name {id, ..} => {
resolver.get_symbol_type(id).ok_or_else(|| "unknown type variable identifer".to_string())
},
ast::ExprKind::Name { id, .. } => resolver
.get_symbol_type(id)
.ok_or_else(|| "unknown type variable identifer".to_string()),
ast::ExprKind::Attribute {value, attr, ..} => {
if let ast::ExprKind::Name {id, ..} = &value.node {
let next_resolver = resolver.get_module_resolver(id).ok_or_else(|| "unknown imported module".to_string())?;
next_resolver.get_symbol_type(attr).ok_or_else(|| "unknown type variable identifer".to_string())
ast::ExprKind::Attribute { value, attr, .. } => {
if let ast::ExprKind::Name { id, .. } = &value.node {
let next_resolver = resolver
.get_module_resolver(id)
.ok_or_else(|| "unknown imported module".to_string())?;
next_resolver
.get_symbol_type(attr)
.ok_or_else(|| "unknown type variable identifer".to_string())
} else {
unimplemented!()
unimplemented!()
// recursively resolve attr thing, FIXME: new problem: how do we handle this?
// # A.py
// class A:
@ -486,8 +495,8 @@ pub fn parse_type_var<T>(input: &ast::Expr<T>, resolver: &dyn SymbolResolver) ->
// class B(Generic[A.A.T]):
// pass
}
},
}
_ => Err("not supported".into())
_ => Err("not supported".into()),
}
}

View File

@ -1,8 +1,11 @@
use crate::typecheck::{
type_inferencer::*,
typedef::{FunSignature, FuncArg, Type, TypeEnum, Unifier},
};
use rustpython_parser::ast;
use rustpython_parser::ast::{Cmpop, Operator, Unaryop};
use std::borrow::Borrow;
use std::collections::HashMap;
use rustpython_parser::ast::{Cmpop, Operator, Unaryop};
use crate::typecheck::{type_inferencer::*, typedef::{FunSignature, FuncArg, TypeEnum, Unifier, Type}};
use rustpython_parser::ast;
pub fn binop_name(op: &Operator) -> &'static str {
match op {
@ -42,206 +45,218 @@ pub fn binop_assign_name(op: &Operator) -> &'static str {
pub fn unaryop_name(op: &Unaryop) -> &'static str {
match op {
Unaryop::UAdd => "__pos__",
Unaryop::USub => "__neg__",
Unaryop::Not => "__not__",
Unaryop::UAdd => "__pos__",
Unaryop::USub => "__neg__",
Unaryop::Not => "__not__",
Unaryop::Invert => "__inv__",
}
}
pub fn comparison_name(op: &Cmpop) -> Option<&'static str> {
match op {
Cmpop::Lt => Some("__lt__"),
Cmpop::LtE => Some("__le__"),
Cmpop::Gt => Some("__gt__"),
Cmpop::GtE => Some("__ge__"),
Cmpop::Eq => Some("__eq__"),
Cmpop::Lt => Some("__lt__"),
Cmpop::LtE => Some("__le__"),
Cmpop::Gt => Some("__gt__"),
Cmpop::GtE => Some("__ge__"),
Cmpop::Eq => Some("__eq__"),
Cmpop::NotEq => Some("__ne__"),
_ => None,
}
}
pub fn impl_binop(unifier: &mut Unifier, _store: &PrimitiveStore, ty: Type, other_ty: &[Type], ret_ty: Type, ops: &[ast::Operator]) {
if let TypeEnum::TObj {fields, ..} = unifier.get_ty(ty).borrow() {
pub fn impl_binop(
unifier: &mut Unifier,
_store: &PrimitiveStore,
ty: Type,
other_ty: &[Type],
ret_ty: Type,
ops: &[ast::Operator],
) {
if let TypeEnum::TObj { fields, .. } = unifier.get_ty(ty).borrow() {
for op in ops {
fields.borrow_mut().insert(
binop_name(op).into(),
{
let other = if other_ty.len() == 1 {
other_ty[0]
} else {
unifier.get_fresh_var_with_range(other_ty).0
};
unifier.add_ty(TypeEnum::TFunc(FunSignature {
ret: ret_ty,
vars: HashMap::new(),
args: vec![FuncArg {
ty: other,
default_value: None,
name: "other".into()
}]
}))
}
);
fields.borrow_mut().insert(binop_name(op).into(), {
let other = if other_ty.len() == 1 {
other_ty[0]
} else {
unifier.get_fresh_var_with_range(other_ty).0
};
unifier.add_ty(TypeEnum::TFunc(FunSignature {
ret: ret_ty,
vars: HashMap::new(),
args: vec![FuncArg { ty: other, default_value: None, name: "other".into() }],
}))
});
fields.borrow_mut().insert(
binop_assign_name(op).into(),
{
let other = if other_ty.len() == 1 {
other_ty[0]
} else {
unifier.get_fresh_var_with_range(other_ty).0
};
unifier.add_ty(TypeEnum::TFunc(FunSignature {
ret: ret_ty,
vars: HashMap::new(),
args: vec![FuncArg {
ty: other,
default_value: None,
name: "other".into()
}]
}))
}
);
fields.borrow_mut().insert(binop_assign_name(op).into(), {
let other = if other_ty.len() == 1 {
other_ty[0]
} else {
unifier.get_fresh_var_with_range(other_ty).0
};
unifier.add_ty(TypeEnum::TFunc(FunSignature {
ret: ret_ty,
vars: HashMap::new(),
args: vec![FuncArg { ty: other, default_value: None, name: "other".into() }],
}))
});
}
} else { unreachable!("") }
} else {
unreachable!("")
}
}
pub fn impl_unaryop(unifier: &mut Unifier, _store: &PrimitiveStore, ty: Type, ret_ty: Type, ops: &[ast::Unaryop]) {
if let TypeEnum::TObj {fields, ..} = unifier.get_ty(ty).borrow() {
pub fn impl_unaryop(
unifier: &mut Unifier,
_store: &PrimitiveStore,
ty: Type,
ret_ty: Type,
ops: &[ast::Unaryop],
) {
if let TypeEnum::TObj { fields, .. } = unifier.get_ty(ty).borrow() {
for op in ops {
fields.borrow_mut().insert(
unaryop_name(op).into(),
unifier.add_ty(TypeEnum::TFunc(FunSignature {
ret: ret_ty,
vars: HashMap::new(),
args: vec![]
}))
args: vec![],
})),
);
}
} else { unreachable!() }
} else {
unreachable!()
}
}
pub fn impl_cmpop(unifier: &mut Unifier, store: &PrimitiveStore, ty: Type, other_ty: Type, ops: &[ast::Cmpop]) {
if let TypeEnum::TObj {fields, ..} = unifier.get_ty(ty).borrow() {
pub fn impl_cmpop(
unifier: &mut Unifier,
store: &PrimitiveStore,
ty: Type,
other_ty: Type,
ops: &[ast::Cmpop],
) {
if let TypeEnum::TObj { fields, .. } = unifier.get_ty(ty).borrow() {
for op in ops {
fields.borrow_mut().insert(
comparison_name(op).unwrap().into(),
unifier.add_ty(TypeEnum::TFunc(FunSignature {
ret: store.bool,
vars: HashMap::new(),
args: vec![FuncArg {
ty: other_ty,
default_value: None,
name: "other".into()
}]
}))
args: vec![FuncArg { ty: other_ty, default_value: None, name: "other".into() }],
})),
);
}
} else { unreachable!() }
} else {
unreachable!()
}
}
/// Add, Sub, Mult, Pow
pub fn impl_basic_arithmetic(unifier: &mut Unifier, store: &PrimitiveStore, ty: Type, other_ty: &[Type], ret_ty: Type) {
impl_binop(unifier, store, ty, other_ty, ret_ty, &[
ast::Operator::Add,
ast::Operator::Sub,
ast::Operator::Mult,
])
pub fn impl_basic_arithmetic(
unifier: &mut Unifier,
store: &PrimitiveStore,
ty: Type,
other_ty: &[Type],
ret_ty: Type,
) {
impl_binop(
unifier,
store,
ty,
other_ty,
ret_ty,
&[ast::Operator::Add, ast::Operator::Sub, ast::Operator::Mult],
)
}
pub fn impl_pow(unifier: &mut Unifier, store: &PrimitiveStore, ty: Type, other_ty: &[Type], ret_ty: Type) {
impl_binop(unifier, store, ty, other_ty, ret_ty, &[
ast::Operator::Pow,
])
pub fn impl_pow(
unifier: &mut Unifier,
store: &PrimitiveStore,
ty: Type,
other_ty: &[Type],
ret_ty: Type,
) {
impl_binop(unifier, store, ty, other_ty, ret_ty, &[ast::Operator::Pow])
}
/// BitOr, BitXor, BitAnd
pub fn impl_bitwise_arithmetic(unifier: &mut Unifier, store: &PrimitiveStore, ty: Type) {
impl_binop(unifier, store, ty, &[ty], ty, &[
ast::Operator::BitAnd,
ast::Operator::BitOr,
ast::Operator::BitXor,
])
impl_binop(
unifier,
store,
ty,
&[ty],
ty,
&[ast::Operator::BitAnd, ast::Operator::BitOr, ast::Operator::BitXor],
)
}
/// LShift, RShift
pub fn impl_bitwise_shift(unifier: &mut Unifier, store: &PrimitiveStore, ty: Type) {
impl_binop(unifier, store, ty, &[ty], ty, &[
ast::Operator::LShift,
ast::Operator::RShift,
])
impl_binop(unifier, store, ty, &[ty], ty, &[ast::Operator::LShift, ast::Operator::RShift])
}
/// Div
pub fn impl_div(unifier: &mut Unifier, store: &PrimitiveStore, ty: Type, other_ty: &[Type]) {
impl_binop(unifier, store, ty, other_ty, store.float, &[
ast::Operator::Div,
])
impl_binop(unifier, store, ty, other_ty, store.float, &[ast::Operator::Div])
}
/// FloorDiv
pub fn impl_floordiv(unifier: &mut Unifier, store: &PrimitiveStore, ty: Type, other_ty: &[Type], ret_ty: Type) {
impl_binop(unifier, store, ty, other_ty, ret_ty, &[
ast::Operator::FloorDiv,
])
pub fn impl_floordiv(
unifier: &mut Unifier,
store: &PrimitiveStore,
ty: Type,
other_ty: &[Type],
ret_ty: Type,
) {
impl_binop(unifier, store, ty, other_ty, ret_ty, &[ast::Operator::FloorDiv])
}
/// Mod
pub fn impl_mod(unifier: &mut Unifier, store: &PrimitiveStore, ty: Type, other_ty: &[Type], ret_ty: Type) {
impl_binop(unifier, store, ty, other_ty, ret_ty, &[
ast::Operator::Mod,
])
pub fn impl_mod(
unifier: &mut Unifier,
store: &PrimitiveStore,
ty: Type,
other_ty: &[Type],
ret_ty: Type,
) {
impl_binop(unifier, store, ty, other_ty, ret_ty, &[ast::Operator::Mod])
}
/// UAdd, USub
pub fn impl_sign(unifier: &mut Unifier, store: &PrimitiveStore, ty: Type) {
impl_unaryop(unifier, store, ty, ty, &[
ast::Unaryop::UAdd,
ast::Unaryop::USub,
])
impl_unaryop(unifier, store, ty, ty, &[ast::Unaryop::UAdd, ast::Unaryop::USub])
}
/// Invert
pub fn impl_invert(unifier: &mut Unifier, store: &PrimitiveStore, ty: Type) {
impl_unaryop(unifier, store, ty, ty, &[
ast::Unaryop::Invert,
])
impl_unaryop(unifier, store, ty, ty, &[ast::Unaryop::Invert])
}
/// Not
pub fn impl_not(unifier: &mut Unifier, store: &PrimitiveStore, ty: Type) {
impl_unaryop(unifier, store, ty, store.bool, &[
ast::Unaryop::Not,
])
impl_unaryop(unifier, store, ty, store.bool, &[ast::Unaryop::Not])
}
/// Lt, LtE, Gt, GtE
pub fn impl_comparison(unifier: &mut Unifier, store: &PrimitiveStore, ty: Type, other_ty: Type) {
impl_cmpop(unifier, store, ty, other_ty, &[
ast::Cmpop::Lt,
ast::Cmpop::Gt,
ast::Cmpop::LtE,
ast::Cmpop::GtE,
])
impl_cmpop(
unifier,
store,
ty,
other_ty,
&[ast::Cmpop::Lt, ast::Cmpop::Gt, ast::Cmpop::LtE, ast::Cmpop::GtE],
)
}
/// Eq, NotEq
pub fn impl_eq(unifier: &mut Unifier, store: &PrimitiveStore, ty: Type) {
impl_cmpop(unifier, store, ty, ty, &[
ast::Cmpop::Eq,
ast::Cmpop::NotEq,
])
impl_cmpop(unifier, store, ty, ty, &[ast::Cmpop::Eq, ast::Cmpop::NotEq])
}
pub fn set_primitives_magic_methods(store: &PrimitiveStore, unifier: &mut Unifier) {
let PrimitiveStore {
int32: int32_t,
int64: int64_t,
float: float_t,
bool: bool_t,
..
} = *store;
let PrimitiveStore { int32: int32_t, int64: int64_t, float: float_t, bool: bool_t, .. } =
*store;
/* int32 ======== */
impl_basic_arithmetic(unifier, store, int32_t, &[int32_t], int32_t);
impl_pow(unifier, store, int32_t, &[int32_t], int32_t);