Compare commits

..
Author SHA1 Message Date
selimaj-dev 872fdf09e6 C like var declerations 2026-04-22 21:03:08 +02:00
selimaj-dev 4118e4b562 Mist 2026-04-22 20:49:29 +02:00
selimaj-dev 753f8501c0 Improving compiler 2026-04-22 20:48:13 +02:00
selimaj-dev 040be1df88 Working variables 2026-04-22 20:28:39 +02:00
selimaj-dev 4f2d1d3664 Working c like parsing 2026-04-22 16:23:59 +02:00
selimaj-dev 2576c77466 Adding mist system 2026-04-22 15:21:42 +02:00
selimaj-dev 1ee3ae317e Improved implementations 2026-04-22 15:16:15 +02:00
selimaj-dev 27a772dad8 Mist Script 2026-04-22 15:03:34 +02:00
selimaj-dev bd262f98f8 If statement expression 2026-04-22 14:38:37 +02:00
selimaj-dev 22cb5aade3 Merge pull request #11 from mist-go/working-go-compiler
Working go compiler
2026-04-22 09:59:38 +02:00
selimaj-dev d514fa9efc Fix expression statement 2026-04-22 09:49:09 +02:00
selimaj-dev 141d2ab7b0 Token naming fix 2026-04-22 09:42:55 +02:00
selimaj-dev ea6aabd896 Working imports 2026-04-22 09:34:23 +02:00
selimaj-dev 53b87eb02e More flexible types 2026-04-22 09:27:29 +02:00
selimaj-dev 797eed173c HIR package importing 2026-04-22 09:23:37 +02:00
selimaj-dev 10a304bfe4 Mist mapping 2026-04-22 09:05:32 +02:00
selimaj-dev 842383ba6e Testing scope importing 2026-04-22 08:56:44 +02:00
selimaj-dev 664e8aba74 Working mapgen 2026-04-22 08:43:34 +02:00
selimaj-dev f5126ad37c mapgen 2026-04-21 18:19:42 +02:00
selimaj-dev 5b5b6aa795 Fixed types 2026-04-21 12:39:55 +02:00
selimaj-dev e45546ad2e Main function 2026-04-21 12:29:06 +02:00
selimaj-dev 18fedba119 Package 2026-04-21 11:28:26 +02:00
selimaj-dev 508b180d63 Compiler 2026-04-21 11:20:20 +02:00
selimaj-dev 1822ce7bfb Working golang codegen 2026-04-20 22:15:37 +02:00
selimaj-dev a28d431a03 Go codegen 2026-04-20 19:54:17 +02:00
selimaj-dev 4d6e749866 Added codegen lib 2026-04-20 19:20:56 +02:00
selimaj-dev 101a8c65ac Merge pull request #10 from mist-go/decl-export
Declration export
2026-04-17 13:06:39 +02:00
selimaj-dev 60d668a222 Field name exporting 2026-04-17 13:03:16 +02:00
selimaj-dev ec30d18e15 field exporting 2026-04-17 12:56:08 +02:00
selimaj-dev d2f7c0c69b field naming 2026-04-17 12:35:20 +02:00
selimaj-dev 7996a3ffaa Decleration identifier 2026-04-17 12:04:44 +02:00
selimaj-dev d0b7898ad7 Identifier fields 2026-04-17 04:06:29 +02:00
selimaj-dev 41f47f3241 Cleaner json output 2026-04-17 03:59:00 +02:00
selimaj-dev e90e9a1f29 Parse with semantics 2026-04-17 03:56:14 +02:00
selimaj-dev 02447bf510 Merge pull request #9 from mist-go/hir-refactor
Refactored HIR Analysis for type checking
2026-04-16 10:33:10 +02:00
selimaj-dev 1edcdade33 Incremental naming 2026-04-16 10:25:47 +02:00
selimaj-dev 2d46afff9a Variable indexing 2026-04-16 10:12:49 +02:00
selimaj-dev 3e1fa5cd18 Working type checker with functions 2026-04-16 10:03:24 +02:00
selimaj-dev 819441bc29 Refactored scoping 2026-04-16 09:59:07 +02:00
selimaj-dev 8ed8eed5de Refactored HIR 2026-04-16 09:45:25 +02:00
selimaj-dev d27b607686 Fixed variable declaration expression 2026-04-16 09:14:18 +02:00
selimaj-dev 1babc77957 Scope postfix ast walking 2026-04-16 09:11:06 +02:00
selimaj-dev bf3783a615 Merge pull request #8 from mist-go/improved-type-checking
Improved type checking
2026-04-16 08:47:24 +02:00
selimaj-dev 874246f8c6 More type checking 2026-04-15 21:34:13 +02:00
selimaj-dev 608ca0916c Fixed naming 2026-04-15 21:18:34 +02:00
selimaj-dev 78738a5faf Merge pull request #7 from mist-go/semantic-scope
Semantic scope
2026-04-15 12:38:16 +02:00
selimaj-dev 09c7f07c95 Variable decleration 2026-04-15 12:36:42 +02:00
selimaj-dev 04087abbdf small refactor 2026-04-15 12:18:40 +02:00
selimaj-dev 88bf07b913 function params 2026-04-15 11:57:45 +02:00
selimaj-dev 0d63528d10 Walking ast 2026-04-15 11:47:36 +02:00
selimaj-dev cb03653e88 Refrencing and LocalScope 2026-04-15 11:18:20 +02:00
selimaj-dev 7ce90478f0 working hir 2026-04-15 10:56:39 +02:00
selimaj-dev 4907533441 HIR 2026-04-15 10:51:55 +02:00
selimaj-dev deeb9350b4 Semantics in check 2026-04-15 09:18:56 +02:00
selimaj-dev 6717be038f Top level scope 2026-04-15 09:16:26 +02:00
selimaj-dev 19629363b0 Top level scope 2026-04-15 08:40:28 +02:00
selimaj-dev 6df7f672d0 Simple semantic analysis structure 2026-04-15 08:32:29 +02:00
selimaj-dev 5de519f561 Merge pull request #6 from mist-go/refactor
Refactor
2026-04-15 08:23:06 +02:00
selimaj-dev 1b3b98458d Added semantic 2026-04-15 06:00:57 +02:00
selimaj-dev 1847b2f282 Removed errors 2026-04-15 04:50:53 +02:00
selimaj-dev b0ffd300d9 Removed compiler and codegen 2026-04-15 04:50:31 +02:00
selimaj-dev f3875a4ec9 Separated parser 2026-04-15 04:46:24 +02:00
selimaj-dev 98608a1883 Fix empty function params 2026-04-14 11:24:04 +02:00
selimaj-dev 34b71c17c3 Merge pull request #4 from mist-go/rewritten-parser
Rewritten parser
2026-04-14 10:41:49 +02:00
28 changed files with 2796 additions and 504 deletions
Generated
+27
View File
@@ -17,6 +17,13 @@ version = "1.0.4"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9330f8b2ff13f34540b44e946ef35111825727b38d33286ef986142615121801"
[[package]]
name = "codegen"
version = "0.1.0"
dependencies = [
"parser",
]
[[package]]
name = "cpufeatures"
version = "0.2.17"
@@ -77,6 +84,17 @@ checksum = "f8ca58f447f06ed17d5fc4043ce1b10dd205e060fb3ce5b979b8ed8e59ff3f79"
[[package]]
name = "mist"
version = "0.1.0"
dependencies = [
"codegen",
"parser",
"semantic",
"serde",
"serde_json",
]
[[package]]
name = "parser"
version = "0.1.0"
dependencies = [
"pest",
"pest_derive",
@@ -145,6 +163,15 @@ dependencies = [
"proc-macro2",
]
[[package]]
name = "semantic"
version = "0.1.0"
dependencies = [
"parser",
"serde",
"serde_json",
]
[[package]]
name = "serde"
version = "1.0.228"
+3 -2
View File
@@ -4,7 +4,8 @@ version = "0.1.0"
edition = "2024"
[dependencies]
pest = "2.8.6"
pest_derive = "2.8.6"
serde = { version = "1.0.228", features = ["serde_derive"] }
serde_json = "1.0.149"
parser = { path = "./parser" }
semantic = { path = "./semantic" }
codegen = { path = "./codegen" }
+248
View File
@@ -0,0 +1,248 @@
# This file is automatically @generated by Cargo.
# It is not intended for manual editing.
version = 4
[[package]]
name = "block-buffer"
version = "0.10.4"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "3078c7629b62d3f0439517fa394996acacc5cbc91c5a20d8c658e77abd503a71"
dependencies = [
"generic-array",
]
[[package]]
name = "cfg-if"
version = "1.0.4"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9330f8b2ff13f34540b44e946ef35111825727b38d33286ef986142615121801"
[[package]]
name = "codegen"
version = "0.1.0"
dependencies = [
"parser",
]
[[package]]
name = "cpufeatures"
version = "0.2.17"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "59ed5838eebb26a2bb2e58f6d5b5316989ae9d08bab10e0e6d103e656d1b0280"
dependencies = [
"libc",
]
[[package]]
name = "crypto-common"
version = "0.1.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "78c8292055d1c1df0cce5d180393dc8cce0abec0a7102adb6c7b1eef6016d60a"
dependencies = [
"generic-array",
"typenum",
]
[[package]]
name = "digest"
version = "0.10.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9ed9a281f7bc9b7576e61468ba615a66a5c8cfdff42420a70aa82701a3b1e292"
dependencies = [
"block-buffer",
"crypto-common",
]
[[package]]
name = "generic-array"
version = "0.14.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "85649ca51fd72272d7821adaf274ad91c288277713d9c18820d8499a7ff69e9a"
dependencies = [
"typenum",
"version_check",
]
[[package]]
name = "itoa"
version = "1.0.18"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "8f42a60cbdf9a97f5d2305f08a87dc4e09308d1276d28c869c684d7777685682"
[[package]]
name = "libc"
version = "0.2.185"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "52ff2c0fe9bc6cb6b14a0592c2ff4fa9ceb83eea9db979b0487cd054946a2b8f"
[[package]]
name = "memchr"
version = "2.8.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f8ca58f447f06ed17d5fc4043ce1b10dd205e060fb3ce5b979b8ed8e59ff3f79"
[[package]]
name = "parser"
version = "0.1.0"
dependencies = [
"pest",
"pest_derive",
"serde",
"serde_json",
]
[[package]]
name = "pest"
version = "2.8.6"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "e0848c601009d37dfa3430c4666e147e49cdcf1b92ecd3e63657d8a5f19da662"
dependencies = [
"memchr",
"ucd-trie",
]
[[package]]
name = "pest_derive"
version = "2.8.6"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "11f486f1ea21e6c10ed15d5a7c77165d0ee443402f0780849d1768e7d9d6fe77"
dependencies = [
"pest",
"pest_generator",
]
[[package]]
name = "pest_generator"
version = "2.8.6"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "8040c4647b13b210a963c1ed407c1ff4fdfa01c31d6d2a098218702e6664f94f"
dependencies = [
"pest",
"pest_meta",
"proc-macro2",
"quote",
"syn",
]
[[package]]
name = "pest_meta"
version = "2.8.6"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "89815c69d36021a140146f26659a81d6c2afa33d216d736dd4be5381a7362220"
dependencies = [
"pest",
"sha2",
]
[[package]]
name = "proc-macro2"
version = "1.0.106"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "8fd00f0bb2e90d81d1044c2b32617f68fcb9fa3bb7640c23e9c748e53fb30934"
dependencies = [
"unicode-ident",
]
[[package]]
name = "quote"
version = "1.0.45"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "41f2619966050689382d2b44f664f4bc593e129785a36d6ee376ddf37259b924"
dependencies = [
"proc-macro2",
]
[[package]]
name = "serde"
version = "1.0.228"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9a8e94ea7f378bd32cbbd37198a4a91436180c5bb472411e48b5ec2e2124ae9e"
dependencies = [
"serde_core",
"serde_derive",
]
[[package]]
name = "serde_core"
version = "1.0.228"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "41d385c7d4ca58e59fc732af25c3983b67ac852c1a25000afe1175de458b67ad"
dependencies = [
"serde_derive",
]
[[package]]
name = "serde_derive"
version = "1.0.228"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d540f220d3187173da220f885ab66608367b6574e925011a9353e4badda91d79"
dependencies = [
"proc-macro2",
"quote",
"syn",
]
[[package]]
name = "serde_json"
version = "1.0.149"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "83fc039473c5595ace860d8c4fafa220ff474b3fc6bfdb4293327f1a37e94d86"
dependencies = [
"itoa",
"memchr",
"serde",
"serde_core",
"zmij",
]
[[package]]
name = "sha2"
version = "0.10.9"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "a7507d819769d01a365ab707794a4084392c824f54a7a6a7862f8c3d0892b283"
dependencies = [
"cfg-if",
"cpufeatures",
"digest",
]
[[package]]
name = "syn"
version = "2.0.117"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "e665b8803e7b1d2a727f4023456bbbbe74da67099c585258af0ad9c5013b9b99"
dependencies = [
"proc-macro2",
"quote",
"unicode-ident",
]
[[package]]
name = "typenum"
version = "1.20.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "40ce102ab67701b8526c123c1bab5cbe42d7040ccfd0f64af1a385808d2f43de"
[[package]]
name = "ucd-trie"
version = "0.1.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "2896d95c02a80c6d6a5d6e953d479f5ddf2dfdb6a244441010e373ac0fb88971"
[[package]]
name = "unicode-ident"
version = "1.0.24"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "e6e4313cd5fcd3dad5cafa179702e2b244f760991f45397d14d4ebf38247da75"
[[package]]
name = "version_check"
version = "0.9.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "0b928f33d975fc6ad9f86c8f283853ad26bdd5b10b7f1542aa2fa15e2289105a"
[[package]]
name = "zmij"
version = "1.0.21"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "b8848ee67ecc8aedbaf3e4122217aff892639231befc6a1b58d29fff4c2cabaa"
+7
View File
@@ -0,0 +1,7 @@
[package]
name = "codegen"
version = "0.1.0"
edition = "2024"
[dependencies]
parser = { path = "../parser" }
+371
View File
@@ -0,0 +1,371 @@
use parser::ast::{BinaryOp, Block, Expression, Postfix, Statement, TopLevel, TypeExpr, VarKind};
pub struct GoCodegen {
output: String,
indent: usize,
}
impl GoCodegen {
pub fn new() -> Self {
Self {
output: String::new(),
indent: 0,
}
}
fn indent_str(&self) -> String {
" ".repeat(self.indent)
}
fn add_indented(&mut self, s: &str) {
self.output.push_str(&format!("{}{}", self.indent_str(), s));
}
fn add(&mut self, s: &str) {
self.output.push_str(s);
}
fn addln(&mut self, s: &str) {
self.add(s);
self.add("\n");
}
fn add_indentedln(&mut self, s: &str) {
self.add_indented(s);
self.add("\n");
}
pub fn generate(&mut self, toplevels: &[TopLevel]) -> String {
for tl in toplevels {
self.generate_toplevel(tl);
}
self.output.clone()
}
fn generate_toplevel(&mut self, tl: &TopLevel) {
match tl {
TopLevel::Import(path) => {
let import_path = path.replace("\"", "");
if import_path.starts_with("./") || import_path.starts_with("/") {
self.addln(&format!("import \"{}\"", import_path));
} else {
self.addln(&format!("import \"{}\"", import_path));
}
self.addln("");
}
TopLevel::StructDecl {
export,
name,
fields,
} => {
let name = if *export { name } else { name };
self.addln(&format!("type {} struct {{", name));
self.indent += 1;
for (field_name, (_, ty)) in &fields.0 {
let go_ty = self.translate_type(ty);
self.addln(&format!("{} {}", field_name, go_ty));
}
self.indent -= 1;
self.addln("}\n");
}
TopLevel::FunctionDecl {
export,
name,
params,
return_type,
body,
} => {
let name = if *export {
format!("{}", name)
} else {
name.clone()
};
let params_str = params
.0
.iter()
.map(|(n, (_, t))| format!("{} {}", n, self.translate_type(t)))
.collect::<Vec<_>>()
.join(", ");
let ret_ty = return_type
.as_ref()
.map(|t| self.translate_type(t))
.unwrap_or_else(|| "".to_string());
if ret_ty.is_empty() {
self.addln(&format!("func {}({}) {{", name, params_str));
} else {
self.addln(&format!("func {}({}) {} {{", name, params_str, ret_ty));
}
self.indent += 1;
self.generate_block(body);
self.indent -= 1;
self.addln("}\n");
}
TopLevel::Package(pkg) => {
self.addln(&format!("package {}", pkg));
self.addln("");
}
}
}
fn translate_type(&self, ty: &TypeExpr) -> String {
match ty {
TypeExpr::Identifier(name) => match name.as_str() {
"int" => "int".to_string(),
"float" | "float64" => "float64".to_string(),
"float32" => "float32".to_string(),
"bool" => "bool".to_string(),
"string" => "string".to_string(),
"byte" => "byte".to_string(),
"rune" => "rune".to_string(),
_ => name.clone(),
},
}
}
fn generate_block(&mut self, block: &Block) {
for stmt in &block.0 {
self.generate_statement(stmt);
}
}
fn generate_statement(&mut self, stmt: &Statement) {
match stmt {
Statement::Expression(expr) => {
self.add_indentedln(&format!("{};", self.generate_expression(expr)));
}
Statement::Block(block) => {
self.add_indentedln("{");
self.indent += 1;
self.generate_block(block);
self.indent -= 1;
self.add_indentedln("}\n");
}
Statement::VarDecl {
kind,
name,
init,
type_,
} => {
let go_kind = match kind {
VarKind::Let | VarKind::Const => "var",
VarKind::Var => "var",
};
let init_expr = init
.as_ref()
.map(|e| format!(" = {}", self.generate_expression(e)))
.unwrap_or_else(|| "".to_string());
let type_expr = type_
.clone()
.map(|t| self.translate_type(&t))
.unwrap_or_default();
self.add_indentedln(&format!(
"{} {} {}{};\n",
go_kind, name, type_expr, init_expr
));
}
Statement::VarAssign { target, value } => {
self.add_indentedln(&format!(
"{} = {};\n",
self.generate_expression(target),
self.generate_expression(value)
));
}
Statement::If {
condition,
then_branch,
else_branch,
} => {
self.add_indented(&format!("if {} ", self.generate_expression(condition)));
self.generate_statement(then_branch);
if let Some(else_br) = else_branch {
self.add_indented("else ");
self.generate_statement(else_br);
}
}
Statement::While { condition, body } => {
self.add_indented(&format!("for {} ", self.generate_expression(condition)));
self.generate_statement(body);
}
Statement::For {
init,
condition,
update,
body,
} => {
let (kind, init_name, init_val) = init;
let init_expr = init_val
.as_ref()
.map(|e| format!(" = {}", self.generate_expression(e)))
.unwrap_or_else(|| "".to_string());
let init_str = format!("{} {}{}", self.var_kind_to_go(kind), init_name, init_expr);
let cond_str = condition
.as_ref()
.map(|e| self.generate_expression(e))
.unwrap_or_else(|| "true".to_string());
let update_str = update
.as_ref()
.map(|s| self.generate_expression(&self.stmt_to_expr(s)))
.unwrap_or_else(|| "".to_string());
self.add_indented(&format!("for {}; {}; {} ", init_str, cond_str, update_str));
self.generate_statement(body);
}
Statement::Return(expr) => {
let ret_val = expr
.as_ref()
.map(|e| self.generate_expression(e))
.unwrap_or_else(|| "".to_string());
self.add_indentedln(&format!("return {};\n", ret_val));
}
Statement::Break => {
self.add_indentedln("break;\n");
}
Statement::Continue => {
self.add_indentedln("continue;\n");
}
}
}
fn stmt_to_expr(&self, stmt: &Statement) -> Expression {
match stmt {
Statement::Expression(e) => e.clone(),
_ => Expression::Identifier(String::new()),
}
}
fn var_kind_to_go(&self, kind: &VarKind) -> String {
match kind {
VarKind::Let | VarKind::Const => "var".to_string(),
VarKind::Var => "var".to_string(),
}
}
fn generate_expression(&self, expr: &Expression) -> String {
match expr {
Expression::Identifier(name) => name.clone(),
Expression::IntLiteral(n) => n.to_string(),
Expression::FloatLiteral(n) => n.to_string(),
Expression::BoolLiteral(b) => b.to_string(),
Expression::StringLiteral(s) => format!("\"{}\"", s),
Expression::Postfix { initial, postfixes } => {
let base = self.generate_expression(initial);
self.apply_postfixes(&base, postfixes)
}
}
}
fn apply_postfixes(&self, base: &str, postfixes: &[Postfix]) -> String {
let mut result = base.to_string();
for postfix in postfixes {
result = match postfix {
Postfix::FieldAccess(field) => format!("{}.{}", result, field),
Postfix::Call(args) => {
let args_str = args
.iter()
.map(|a| self.generate_expression(a))
.collect::<Vec<_>>()
.join(", ");
format!("{}({})", result, args_str)
}
Postfix::Index(idx) => format!("{}[{}]", result, self.generate_expression(idx)),
Postfix::Binary(op, rhs) => {
let op_str = match op {
BinaryOp::Plus => "+",
BinaryOp::Minus => "-",
BinaryOp::Multiply => "*",
BinaryOp::Divide => "/",
BinaryOp::Modulo => "%",
BinaryOp::Equal => "==",
BinaryOp::NotEqual => "!=",
BinaryOp::LessThan => "<",
BinaryOp::GreaterThan => ">",
BinaryOp::LessThanOrEqual => "<=",
BinaryOp::GreaterThanOrEqual => ">=",
};
format!("{} {} {}", result, op_str, self.generate_expression(rhs))
}
};
}
result
}
}
impl Default for GoCodegen {
fn default() -> Self {
Self::new()
}
}
#[cfg(test)]
mod tests {
use super::*;
use parser::ast::{Block, Expression, ParamList, Statement, TopLevel, TypeExpr};
use std::collections::HashMap;
#[test]
fn test_int_literal() {
let cg = GoCodegen::new();
let expr = Expression::IntLiteral(42);
let result = cg.generate_expression(&expr);
assert_eq!(result, "42");
}
#[test]
fn test_string_literal() {
let cg = GoCodegen::new();
let expr = Expression::StringLiteral("hello".to_string());
let result = cg.generate_expression(&expr);
assert_eq!(result, "\"hello\"");
}
#[test]
fn test_function_decl() {
let mut cg = GoCodegen::new();
let toplevel = TopLevel::FunctionDecl {
export: true,
name: "main".to_string(),
params: ParamList(HashMap::new()),
return_type: None,
body: Block(vec![Statement::Return(None)]),
};
cg.generate_toplevel(&toplevel);
let output = cg.output.clone();
assert!(output.contains("func main()"));
}
#[test]
fn test_struct_decl() {
let mut cg = GoCodegen::new();
let mut fields = HashMap::new();
fields.insert(
"x".to_string(),
(true, TypeExpr::Identifier("int".to_string())),
);
let toplevel = TopLevel::StructDecl {
export: true,
name: "Point".to_string(),
fields: ParamList(fields),
};
cg.generate_toplevel(&toplevel);
let output = cg.output.clone();
assert!(output.contains("type Point struct"));
assert!(output.contains("x int"));
}
#[test]
fn test_if_statement() {
let mut cg = GoCodegen::new();
let stmt = Statement::If {
condition: Expression::Identifier("x".to_string()),
then_branch: Box::new(Statement::Return(Some(Expression::IntLiteral(1)))),
else_branch: None,
};
cg.generate_statement(&stmt);
let output = cg.output.clone();
assert!(output.contains("if x"));
}
}
+9
View File
@@ -0,0 +1,9 @@
module mapgen
go 1.25.5
require (
golang.org/x/mod v0.35.0 // indirect
golang.org/x/sync v0.20.0 // indirect
golang.org/x/tools v0.44.0 // indirect
)
+6
View File
@@ -0,0 +1,6 @@
golang.org/x/mod v0.35.0 h1:Ww1D637e6Pg+Zb2KrWfHQUnH2dQRLBQyAtpr/haaJeM=
golang.org/x/mod v0.35.0/go.mod h1:+GwiRhIInF8wPm+4AoT6L0FA1QWAad3OMdTRx4tFYlU=
golang.org/x/sync v0.20.0 h1:e0PTpb7pjO8GAtTs2dQ6jYa5BWYlMuX047Dco/pItO4=
golang.org/x/sync v0.20.0/go.mod h1:9xrNwdLfx4jkKbNva9FpL6vEN7evnE43NNNJQ2LF3+0=
golang.org/x/tools v0.44.0 h1:UP4ajHPIcuMjT1GqzDWRlalUEoY+uzoZKnhOjbIPD2c=
golang.org/x/tools v0.44.0/go.mod h1:KA0AfVErSdxRZIsOVipbv3rQhVXTnlU6UhKxHd1seDI=
+188
View File
@@ -0,0 +1,188 @@
package main
import (
"encoding/json"
"fmt"
"go/token"
"go/types"
"os"
"unicode"
"golang.org/x/tools/go/packages"
)
type TypeSymbol string
type VarSymbol struct {
Export bool `json:"export"`
VarType TypeSymbol `json:"var_type"`
Name string `json:"name"`
}
type FunctionSymbol struct {
Export bool `json:"export"`
Name string `json:"name"`
Params map[string]VarSymbol `json:"params"`
ReturnTypes []TypeSymbol `json:"return_types"`
}
type StructSymbol struct {
Export bool `json:"export"`
Name string `json:"name"`
Fields map[string]VarSymbol `json:"fields"`
Methods map[string]FunctionSymbol `json:"methods"`
}
type TopLevelSymbolScope struct {
PackageName string `json:"package_name"`
Structs map[string]StructSymbol `json:"structs"`
Functions map[string]FunctionSymbol `json:"functions"`
}
func main() {
cfg := &packages.Config{
Mode: packages.NeedName |
packages.NeedTypes |
packages.NeedTypesInfo |
packages.NeedModule,
}
// you can expand this to multiple packages later
pkgs, err := packages.Load(cfg, "fmt")
if err != nil {
panic(err)
}
if packages.PrintErrors(pkgs) > 0 {
os.Exit(1)
}
// 🔑 FINAL RESULT: map[import_path]scope
result := make(map[string]TopLevelSymbolScope)
for _, pkg := range pkgs {
scopeData := TopLevelSymbolScope{
PackageName: pkg.Name, // e.g. "http"
Structs: make(map[string]StructSymbol),
Functions: make(map[string]FunctionSymbol),
}
scope := pkg.Types.Scope()
for _, name := range scope.Names() {
if !token.IsExported(name) {
continue
}
obj := scope.Lookup(name)
switch obj := obj.(type) {
case *types.Func:
scopeData.Functions[LowerFirst(name)] = buildFunction(obj)
case *types.TypeName:
if strct, ok := buildStruct(obj); ok {
scopeData.Structs[name] = strct
}
}
}
// 🔑 key = import path (e.g. "net/http")
result[pkg.PkgPath] = scopeData
}
out, err := json.MarshalIndent(result, "", " ")
if err != nil {
panic(err)
}
fmt.Println(string(out))
}
func LowerFirst(s string) string {
if s == "" {
return ""
}
r := []rune(s)
r[0] = unicode.ToLower(r[0])
return string(r)
}
func buildFunction(fn *types.Func) FunctionSymbol {
sig := fn.Type().(*types.Signature)
params := make(map[string]VarSymbol)
for i := 0; i < sig.Params().Len(); i++ {
p := sig.Params().At(i)
name := p.Name()
if name == "" {
name = fmt.Sprintf("param%d", i)
}
params[name] = VarSymbol{
Export: true,
Name: name,
VarType: TypeSymbol(p.Type().String()),
}
}
var returnTypes []TypeSymbol
for i := 0; i < sig.Results().Len(); i++ {
r := sig.Results().At(i)
returnTypes = append(returnTypes, TypeSymbol(r.Type().String()))
}
return FunctionSymbol{
Export: true, // already filtered before calling
Name: fn.Name(),
Params: params,
ReturnTypes: returnTypes,
}
}
func buildStruct(tn *types.TypeName) (StructSymbol, bool) {
underlying := tn.Type().Underlying()
strct, ok := underlying.(*types.Struct)
if !ok {
return StructSymbol{}, false
}
fields := make(map[string]VarSymbol)
for i := 0; i < strct.NumFields(); i++ {
f := strct.Field(i)
if !token.IsExported(f.Name()) {
continue
}
fields[f.Name()] = VarSymbol{
Export: true,
Name: f.Name(),
VarType: TypeSymbol(f.Type().String()),
}
}
methods := make(map[string]FunctionSymbol)
methodSet := types.NewMethodSet(tn.Type())
for i := 0; i < methodSet.Len(); i++ {
m := methodSet.At(i).Obj().(*types.Func)
if !token.IsExported(m.Name()) {
continue
}
methods[LowerFirst(m.Name())] = buildFunction(m)
}
return StructSymbol{
Export: true,
Name: tn.Name(),
Fields: fields,
Methods: methods,
}, true
}
+10
View File
@@ -0,0 +1,10 @@
[package]
name = "parser"
version = "0.1.0"
edition = "2024"
[dependencies]
pest = "2.8.6"
pest_derive = "2.8.6"
serde = { version = "1.0.228", features = ["serde_derive"] }
serde_json = "1.0.149"
+5 -6
View File
@@ -1,13 +1,14 @@
use std::collections::HashMap;
use serde::Serialize;
#[derive(Debug, Clone, Serialize)]
pub struct ParamList(pub Vec<(String, TypeExpr)>);
pub struct ParamList(pub HashMap<String, (bool, TypeExpr)>);
#[derive(Debug, Clone, Serialize)]
pub struct Block(pub Vec<Statement>);
#[derive(Debug, Clone, Serialize)]
#[serde(tag = "type", content = "value")]
pub enum TypeExpr {
Identifier(String),
}
@@ -28,9 +29,9 @@ pub enum BinaryOp {
}
#[derive(Debug, Clone, Serialize)]
#[serde(tag = "type", content = "value")]
pub enum TopLevel {
Import(String),
Package(String),
StructDecl {
export: bool,
name: String,
@@ -46,7 +47,6 @@ pub enum TopLevel {
}
#[derive(Debug, Clone, Serialize)]
#[serde(tag = "type", content = "value")]
pub enum Postfix {
FieldAccess(String),
Call(Vec<Expression>),
@@ -55,7 +55,6 @@ pub enum Postfix {
}
#[derive(Debug, Clone, Serialize)]
#[serde(tag = "type", content = "value")]
pub enum Statement {
// expr;
Expression(Expression),
@@ -68,6 +67,7 @@ pub enum Statement {
kind: VarKind,
name: String,
init: Option<Expression>,
type_: Option<TypeExpr>,
},
VarAssign {
@@ -104,7 +104,6 @@ pub enum Statement {
}
#[derive(Debug, Clone, Serialize)]
#[serde(tag = "type", content = "value")]
pub enum Expression {
Identifier(String),
IntLiteral(i64),
@@ -13,21 +13,23 @@ inner_str = @{ (!"\"" ~ ANY)* }
// ================= TOP LEVEL =================
top_level = _{ import | function_decl | struct_decl }
top_level = _{ import | function_decl | struct_decl | package }
export = { "export" }
package = { "package" ~ identifier ~ ";" }
export = { "public" }
import = { "import" ~ string_lit ~ ";" }
param_list = { param ~ ("," ~ param)* ~ ","? }
param = { identifier ~ ":" ~ type_expr }
param = { export? ~ type_expr ~ identifier }
struct_decl = {
export? ~ "struct" ~ identifier ~ "{" ~ param_list? ~ "}"
}
function_decl = {
export? ~ "function" ~ identifier ~ "(" ~ param_list? ~ ")" ~ (":" ~ type_expr)? ~ block
export? ~ type_expr ~ identifier ~ "(" ~ param_list? ~ ")" ~ block
}
// ================= STATEMENTS =================
@@ -61,10 +63,10 @@ continue_stmt = { "continue" ~ ";" }
// --- variable declarations ---
var_decl_kind = { "let" | "const" | "var" }
// var_decl_kind = { "let" | "const" | "var" }
var_decl = {
var_decl_kind ~ identifier ~ ("=" ~ expr)? ~ ";"
type_expr ~ identifier ~ ("=" ~ expr)? ~ ";"
}
var_assign = {
+391
View File
@@ -0,0 +1,391 @@
use std::collections::HashMap;
use pest::Parser;
use pest_derive::Parser;
pub mod ast;
pub mod script_parser;
use ast::*;
#[derive(Parser)]
#[grammar = "./src/grammar.pest"]
pub struct MistParser;
// convenience alias for pest errors
pub type ParseError = pest::error::Error<Rule>;
pub fn parse(source: &str) -> Result<Vec<TopLevel>, ParseError> {
let mut pairs = MistParser::parse(Rule::program, source)?;
let mut statements = vec![];
for pair in pairs.next().unwrap().into_inner() {
if let Ok(stmt) = TopLevel::try_from(pair) {
statements.push(stmt);
}
}
Ok(statements)
}
impl TryFrom<pest::iterators::Pair<'_, Rule>> for TypeExpr {
type Error = ();
fn try_from(pair: pest::iterators::Pair<'_, Rule>) -> Result<Self, Self::Error> {
if pair.as_str() == "void" {
return Err(());
}
Ok(match pair.as_rule() {
Rule::type_expr => {
let inner = pair.into_inner().next().unwrap();
TypeExpr::try_from(inner)?
}
Rule::identifier => TypeExpr::Identifier(pair.as_str().to_string()),
_ => unimplemented!("{pair:#?}"),
})
}
}
impl From<(bool, pest::iterators::Pair<'_, Rule>)> for ParamList {
fn from((export_allowed, pair): (bool, pest::iterators::Pair<Rule>)) -> Self {
let params = pair
.into_inner()
.map(|p| {
let mut param_inner = p.into_inner();
let export =
if export_allowed && param_inner.peek().unwrap().as_rule() == Rule::export {
param_inner.next().unwrap();
true
} else {
false
};
let param_type = TypeExpr::try_from(param_inner.next().unwrap()).unwrap();
let param_name = param_inner.next().unwrap().as_str().to_string();
(param_name, (export, param_type))
})
.collect();
ParamList(params)
}
}
impl TryFrom<pest::iterators::Pair<'_, Rule>> for TopLevel {
type Error = ();
fn try_from(pair: pest::iterators::Pair<Rule>) -> Result<Self, ()> {
match pair.as_rule() {
Rule::import => {
let path = pair.into_inner().next().unwrap().as_str().to_string();
Ok(TopLevel::Import(path))
}
Rule::package => {
let name = pair.into_inner().next().unwrap().as_str().to_string();
Ok(TopLevel::Package(name))
}
Rule::function_decl => {
let mut inner = pair.into_inner();
let export = if let Some(first) = inner.peek() {
if first.as_rule() == Rule::export {
inner.next();
true
} else {
false
}
} else {
false
};
let return_type = if let Some(next) = inner.peek() {
if next.as_rule() == Rule::type_expr {
TypeExpr::try_from(inner.next().unwrap()).ok()
} else {
None
}
} else {
None
};
let name = inner.next().unwrap().as_str().to_string();
let params = if inner.peek().unwrap().as_rule() == Rule::param_list {
ParamList::from((false, inner.next().unwrap()))
} else {
ParamList(HashMap::new())
};
let body = Block::from(inner.next().unwrap());
Ok(TopLevel::FunctionDecl {
export,
name,
params,
return_type,
body,
})
}
Rule::struct_decl => {
let mut inner = pair.into_inner();
let export = if let Some(first) = inner.peek() {
if first.as_rule() == Rule::export {
inner.next();
true
} else {
false
}
} else {
false
};
let name = inner.next().unwrap().as_str().to_string();
let fields_pair = inner.next().unwrap();
let fields = ParamList::from((true, fields_pair));
Ok(TopLevel::StructDecl {
export,
name,
fields,
})
}
Rule::EOI => Err(()),
_ => unimplemented!("TopLevel parsing not implemented yet {:?}", pair.as_rule()),
}
}
}
impl From<pest::iterators::Pair<'_, Rule>> for Block {
fn from(pair: pest::iterators::Pair<Rule>) -> Self {
let statements = pair
.into_inner()
.flat_map(|pair| {
if pair.as_rule() == Rule::statement_list {
pair.into_inner().map(Statement::from).collect()
} else {
vec![Statement::from(pair)]
}
})
.collect();
Block(statements)
}
}
impl From<pest::iterators::Pair<'_, Rule>> for Statement {
fn from(pair: pest::iterators::Pair<Rule>) -> Self {
match pair.as_rule() {
Rule::statement => {
let inner = pair.into_inner().next().unwrap();
Statement::from(inner)
}
Rule::expr_stmt => {
let expr_pair = pair.into_inner().next().unwrap();
Statement::Expression(Expression::from(expr_pair))
}
Rule::block => Statement::Block(Block::from(pair.into_inner().next().unwrap())),
Rule::var_decl => {
let mut inner = pair.into_inner();
let type_ = inner.next().map(TypeExpr::try_from).unwrap().ok();
let name = inner.next().unwrap().as_str().to_string();
let init = inner.next().map(Expression::from);
Statement::VarDecl {
kind: VarKind::Var,
name: name.as_str().to_string(),
init,
type_,
}
}
Rule::return_stmt => {
let mut inner = pair.into_inner();
let expr = inner.next().map(Expression::from);
Statement::Return(expr)
}
Rule::break_stmt => Statement::Break,
Rule::continue_stmt => Statement::Continue,
Rule::if_stmt => {
let mut inner = pair.into_inner();
let condition = Expression::from(inner.next().unwrap());
let then_branch = Statement::from(inner.next().unwrap());
let else_branch = inner.next().map(Statement::from);
Statement::If {
condition,
then_branch: Box::new(then_branch),
else_branch: else_branch.map(Box::new),
}
}
Rule::while_stmt => {
let mut inner = pair.into_inner();
let condition = Expression::from(inner.next().unwrap());
let body = Statement::from(inner.next().unwrap());
Statement::While {
condition,
body: Box::new(body),
}
}
Rule::for_stmt => {
let mut inner = pair.into_inner();
let init = inner
.next()
.map(|p| match p.as_rule() {
Rule::var_decl => {
let mut it = p.into_inner();
let kind = match it.next().unwrap().as_str() {
"let" => VarKind::Let,
"const" => VarKind::Const,
"var" => VarKind::Var,
_ => unreachable!(),
};
let name = it.next().unwrap().as_str().to_string();
let init_expr = it
.next()
.map(|e| Expression::from(e.into_inner().next().unwrap()));
(kind, name, init_expr)
}
_ => unimplemented!(
"For loop init parsing not implemented yet: {:?}",
p.as_rule()
),
})
.unwrap();
let condition = inner.next().map(Expression::from);
let update = inner.next().map(parse_var_assign_no_semicolon);
let body = Statement::from(inner.next().unwrap());
Statement::For {
init,
condition,
update: update.map(Box::new),
body: Box::new(body),
}
}
Rule::var_assign => {
let mut inner = pair.into_inner();
let target = Expression::from(inner.next().unwrap());
let value = Expression::from(inner.next().unwrap());
Statement::VarAssign { target, value }
}
_ => unimplemented!(
"Statement parsing not implemented yet: {:?}",
pair.as_rule()
),
}
}
}
impl From<pest::iterators::Pair<'_, Rule>> for Expression {
fn from(pair: pest::iterators::Pair<Rule>) -> Self {
match pair.as_rule() {
Rule::expr => {
let mut inner = pair.into_inner();
let exp = Expression::from(inner.next().unwrap());
if inner.len() > 0 {
Expression::Postfix {
initial: Box::new(exp),
postfixes: inner.map(|p| Postfix::from(p)).collect(),
}
} else {
exp
}
}
Rule::primary => Expression::from(pair.into_inner().next().unwrap()),
Rule::identifier => Expression::Identifier(pair.as_str().to_string()),
Rule::integer => {
let value = pair.as_str().parse::<i64>().unwrap();
Expression::IntLiteral(value)
}
Rule::float => {
let value = pair.as_str().parse::<f64>().unwrap();
Expression::FloatLiteral(value)
}
Rule::boolean => {
let value = pair.as_str().parse::<bool>().unwrap();
Expression::BoolLiteral(value)
}
Rule::string_lit => {
let inner_str = pair.into_inner().next().unwrap().as_str();
Expression::StringLiteral(inner_str.to_string())
}
_ => unimplemented!(
"Expression parsing not implemented yet {:?}",
pair.as_rule()
),
}
}
}
impl From<pest::iterators::Pair<'_, Rule>> for Postfix {
fn from(pair: pest::iterators::Pair<Rule>) -> Self {
match pair.as_rule() {
Rule::postfix => Postfix::from(pair.into_inner().next().unwrap()),
Rule::field_px => {
let field_name = pair.into_inner().next().unwrap().as_str().to_string();
Postfix::FieldAccess(field_name)
}
Rule::call_px => Postfix::Call(pair.into_inner().map(Expression::from).collect()),
Rule::index_px => Postfix::Index(Expression::from(pair.into_inner().next().unwrap())),
Rule::binary_px => {
let mut inner = pair.into_inner();
let op_pair = inner.next().unwrap();
let op = match op_pair.as_str() {
"+" => BinaryOp::Plus,
"-" => BinaryOp::Minus,
"*" => BinaryOp::Multiply,
"/" => BinaryOp::Divide,
"%" => BinaryOp::Modulo,
"==" => BinaryOp::Equal,
"!=" => BinaryOp::NotEqual,
"<" => BinaryOp::LessThan,
">" => BinaryOp::GreaterThan,
"<=" => BinaryOp::LessThanOrEqual,
">=" => BinaryOp::GreaterThanOrEqual,
_ => {
unimplemented!("Binary operator not implemented yet: {}", op_pair.as_str())
}
};
Postfix::Binary(op, Expression::from(inner.next().unwrap()))
}
_ => unimplemented!("Postfix parsing not implemented yet {:?}", pair.as_rule()),
}
}
}
fn parse_var_assign_no_semicolon(pair: pest::iterators::Pair<Rule>) -> Statement {
let mut inner = pair.into_inner();
let target = Expression::from(inner.next().unwrap());
let value = Expression::from(inner.next().unwrap());
Statement::VarAssign { target, value }
}
+120
View File
@@ -0,0 +1,120 @@
WHITESPACE = _{ " " | "\t" | "\r" | "\n" }
COMMENT = _{ "//" ~ (!"\n" ~ ANY)* }
program = { SOI ~ top_level* ~ EOI }
// primitives
identifier = @{ (ASCII_ALPHA | "_") ~ (ASCII_ALPHANUMERIC | "_")* }
integer = @{ ASCII_DIGIT+ }
float = @{ ASCII_DIGIT+ ~ "." ~ ASCII_DIGIT+ }
boolean = { "true" | "false" }
string_lit = { "\"" ~ inner_str ~ "\"" }
inner_str = @{ (!"\"" ~ ANY)* }
// ================= TOP LEVEL =================
top_level = _{ import | function_decl | struct_decl | package }
package = { "package" ~ identifier ~ ";" }
export = { "export" }
import = { "import" ~ string_lit ~ ";" }
param_list = { param ~ ("," ~ param)* ~ ","? }
param = { export? ~ identifier ~ ":" ~ type_expr }
struct_decl = {
export? ~ "struct" ~ identifier ~ "{" ~ param_list? ~ "}"
}
function_decl = {
export? ~ "function" ~ identifier ~ "(" ~ param_list? ~ ")" ~ (":" ~ type_expr)? ~ block
}
// ================= STATEMENTS =================
block = { "{" ~ statement_list ~ "}" }
statement_list = { statement* }
statement = _{
expr_stmt
| var_decl
| var_assign
| if_stmt
| while_stmt
| for_stmt
| return_stmt
| break_stmt
| continue_stmt
| block
}
// --- basic statements ---
expr_stmt = { expr ~ ";" }
return_stmt = { "return" ~ expr? ~ ";" }
break_stmt = { "break" ~ ";" }
continue_stmt = { "continue" ~ ";" }
// --- variable declarations ---
var_decl_kind = { "let" | "const" | "var" }
var_decl = {
var_decl_kind ~ identifier ~ ("=" ~ expr)? ~ ";"
}
var_assign = {
expr ~ "=" ~ expr ~ ";"
}
// --- control flow ---
if_stmt = {
"if" ~ "(" ~ expr ~ ")" ~ statement ~ ("else" ~ statement)?
}
while_stmt = {
"while" ~ "(" ~ expr ~ ")" ~ statement
}
for_stmt = {
"for" ~ "(" ~ var_decl ~ expr ~ ";" ~ var_assign_no_semicolon ~ ")" ~ statement
}
var_assign_no_semicolon = {
expr ~ "=" ~ expr
}
// ================= EXPRESSIONS =================
type_expr = { identifier }
primary = {
integer
| float
| boolean
| string_lit
| identifier
}
postfix = {
field_px
| call_px
| index_px
| binary_px
}
field_px = { "." ~ identifier }
call_px = { "(" ~ (expr ~ ("," ~ expr)*)? ~ ")" }
index_px = { "[" ~ expr ~ "]" }
binary_px = { bin_op ~ expr }
bin_op = { "+" | "-" | "*" | "/" | "%" | "==" | "!=" | "<" | ">" | "<=" | ">=" }
expr = {
primary ~ postfix*
}
@@ -1,23 +1,24 @@
use std::collections::HashMap;
use pest::Parser;
use pest_derive::Parser;
pub mod ast;
use ast::*;
use crate::ast::*;
#[derive(Parser)]
#[grammar = "./src/parser/grammar.pest"]
pub struct MistParser;
#[grammar = "./src/script_grammar.pest"]
pub struct MistScriptParser;
// convenience alias for pest errors
pub type ParseError = pest::error::Error<Rule>;
pub fn parse(source: &str) -> Result<Vec<TopLevel>, ParseError> {
let mut pairs = MistParser::parse(Rule::program, source)?;
let mut pairs = MistScriptParser::parse(Rule::program, source)?;
let mut statements = vec![];
for pair in pairs.next().unwrap().into_inner() {
if let Some(stmt) = TopLevel::from_pair(pair) {
if let Ok(stmt) = TopLevel::try_from(pair) {
statements.push(stmt);
}
}
@@ -25,12 +26,12 @@ pub fn parse(source: &str) -> Result<Vec<TopLevel>, ParseError> {
Ok(statements)
}
impl TypeExpr {
pub fn from_pair(pair: pest::iterators::Pair<Rule>) -> Self {
impl From<pest::iterators::Pair<'_, Rule>> for TypeExpr {
fn from(pair: pest::iterators::Pair<Rule>) -> Self {
match pair.as_rule() {
Rule::type_expr => {
let inner = pair.into_inner().next().unwrap();
TypeExpr::from_pair(inner)
TypeExpr::from(inner)
}
Rule::identifier => TypeExpr::Identifier(pair.as_str().to_string()),
_ => unimplemented!("TypeExpr parsing not implemented yet"),
@@ -38,15 +39,22 @@ impl TypeExpr {
}
}
impl ParamList {
pub fn from_pair(pair: pest::iterators::Pair<Rule>) -> Self {
impl From<(bool, pest::iterators::Pair<'_, Rule>)> for ParamList {
fn from((export_allowed, pair): (bool, pest::iterators::Pair<Rule>)) -> Self {
let params = pair
.into_inner()
.map(|p| {
let mut param_inner = p.into_inner();
let export =
if export_allowed && param_inner.peek().unwrap().as_rule() == Rule::export {
param_inner.next().unwrap();
true
} else {
false
};
let param_name = param_inner.next().unwrap().as_str().to_string();
let param_type = TypeExpr::from_pair(param_inner.next().unwrap());
(param_name, param_type)
let param_type = TypeExpr::from(param_inner.next().unwrap());
(param_name, (export, param_type))
})
.collect();
@@ -54,12 +62,17 @@ impl ParamList {
}
}
impl TopLevel {
pub fn from_pair(pair: pest::iterators::Pair<Rule>) -> Option<Self> {
impl TryFrom<pest::iterators::Pair<'_, Rule>> for TopLevel {
type Error = ();
fn try_from(pair: pest::iterators::Pair<Rule>) -> Result<Self, ()> {
match pair.as_rule() {
Rule::import => {
let path = pair.into_inner().next().unwrap().as_str().to_string();
Some(TopLevel::Import(path))
Ok(TopLevel::Import(path))
}
Rule::package => {
let name = pair.into_inner().next().unwrap().as_str().to_string();
Ok(TopLevel::Package(name))
}
Rule::function_decl => {
let mut inner = pair.into_inner();
@@ -75,15 +88,14 @@ impl TopLevel {
false
};
let name = inner.next().unwrap().as_str().to_string();
let params_pair = inner.next().unwrap();
let params = if params_pair.as_rule() == Rule::param_list {
ParamList::from_pair(params_pair)
let params = if inner.peek().unwrap().as_rule() == Rule::param_list {
ParamList::from((false, inner.next().unwrap()))
} else {
ParamList(vec![])
ParamList(HashMap::new())
};
let return_type = if let Some(next) = inner.peek() {
if next.as_rule() == Rule::type_expr {
Some(TypeExpr::from_pair(inner.next().unwrap()))
Some(TypeExpr::from(inner.next().unwrap()))
} else {
None
}
@@ -91,9 +103,9 @@ impl TopLevel {
None
};
let body = Block::from_pair(inner.next().unwrap());
let body = Block::from(inner.next().unwrap());
Some(TopLevel::FunctionDecl {
Ok(TopLevel::FunctionDecl {
export,
name,
params,
@@ -116,30 +128,30 @@ impl TopLevel {
};
let name = inner.next().unwrap().as_str().to_string();
let fields_pair = inner.next().unwrap();
let fields = ParamList::from_pair(fields_pair);
let fields = ParamList::from((true, fields_pair));
Some(TopLevel::StructDecl {
Ok(TopLevel::StructDecl {
export,
name,
fields,
})
}
Rule::EOI => None,
Rule::EOI => Err(()),
_ => unimplemented!("TopLevel parsing not implemented yet {:?}", pair.as_rule()),
}
}
}
impl Block {
pub fn from_pair(pair: pest::iterators::Pair<Rule>) -> Self {
impl From<pest::iterators::Pair<'_, Rule>> for Block {
fn from(pair: pest::iterators::Pair<Rule>) -> Self {
let statements = pair
.into_inner()
.flat_map(|pair| {
if pair.as_rule() == Rule::statement_list {
pair.into_inner().map(Statement::from_pair).collect()
pair.into_inner().map(Statement::from).collect()
} else {
vec![Statement::from_pair(pair)]
vec![Statement::from(pair)]
}
})
.collect();
@@ -147,20 +159,20 @@ impl Block {
}
}
impl Statement {
pub fn from_pair(pair: pest::iterators::Pair<Rule>) -> Self {
impl From<pest::iterators::Pair<'_, Rule>> for Statement {
fn from(pair: pest::iterators::Pair<Rule>) -> Self {
match pair.as_rule() {
Rule::statement => {
let inner = pair.into_inner().next().unwrap();
Statement::from_pair(inner)
Statement::from(inner)
}
Rule::expr_stmt => {
let expr_pair = pair.into_inner().next().unwrap();
Statement::Expression(Expression::from_pair(expr_pair))
Statement::Expression(Expression::from(expr_pair))
}
Rule::block => Statement::Block(Block::from_pair(pair.into_inner().next().unwrap())),
Rule::block => Statement::Block(Block::from(pair.into_inner().next().unwrap())),
Rule::var_decl => {
let mut inner = pair.into_inner();
@@ -168,10 +180,7 @@ impl Statement {
let kind_pair = inner.next().unwrap(); // let/const/var
let name_pair = inner.next().unwrap(); // identifier
let init = inner.next().map(|expr_pair| {
// expects "=" expr
Expression::from_pair(expr_pair.into_inner().next().unwrap())
});
let init = inner.next().map(Expression::from);
let kind = match kind_pair.as_str() {
"let" => VarKind::Let,
@@ -184,13 +193,14 @@ impl Statement {
kind,
name: name_pair.as_str().to_string(),
init,
type_: None,
}
}
Rule::return_stmt => {
let mut inner = pair.into_inner();
let expr = inner.next().map(Expression::from_pair);
let expr = inner.next().map(Expression::from);
Statement::Return(expr)
}
@@ -202,10 +212,10 @@ impl Statement {
Rule::if_stmt => {
let mut inner = pair.into_inner();
let condition = Expression::from_pair(inner.next().unwrap());
let then_branch = Statement::from_pair(inner.next().unwrap());
let condition = Expression::from(inner.next().unwrap());
let then_branch = Statement::from(inner.next().unwrap());
let else_branch = inner.next().map(Statement::from_pair);
let else_branch = inner.next().map(Statement::from);
Statement::If {
condition,
@@ -217,8 +227,8 @@ impl Statement {
Rule::while_stmt => {
let mut inner = pair.into_inner();
let condition = Expression::from_pair(inner.next().unwrap());
let body = Statement::from_pair(inner.next().unwrap());
let condition = Expression::from(inner.next().unwrap());
let body = Statement::from(inner.next().unwrap());
Statement::While {
condition,
@@ -245,7 +255,7 @@ impl Statement {
let name = it.next().unwrap().as_str().to_string();
let init_expr = it
.next()
.map(|e| Expression::from_pair(e.into_inner().next().unwrap()));
.map(|e| Expression::from(e.into_inner().next().unwrap()));
(kind, name, init_expr)
}
@@ -256,9 +266,9 @@ impl Statement {
})
.unwrap();
let condition = inner.next().map(Expression::from_pair);
let condition = inner.next().map(Expression::from);
let update = inner.next().map(parse_var_assign_no_semicolon);
let body = Statement::from_pair(inner.next().unwrap());
let body = Statement::from(inner.next().unwrap());
Statement::For {
init,
@@ -270,8 +280,8 @@ impl Statement {
Rule::var_assign => {
let mut inner = pair.into_inner();
let target = Expression::from_pair(inner.next().unwrap());
let value = Expression::from_pair(inner.next().unwrap());
let target = Expression::from(inner.next().unwrap());
let value = Expression::from(inner.next().unwrap());
Statement::VarAssign { target, value }
}
@@ -284,23 +294,23 @@ impl Statement {
}
}
impl Expression {
pub fn from_pair(pair: pest::iterators::Pair<Rule>) -> Self {
impl From<pest::iterators::Pair<'_, Rule>> for Expression {
fn from(pair: pest::iterators::Pair<Rule>) -> Self {
match pair.as_rule() {
Rule::expr => {
let mut inner = pair.into_inner();
let exp = Expression::from_pair(inner.next().unwrap());
let exp = Expression::from(inner.next().unwrap());
if inner.len() > 0 {
Expression::Postfix {
initial: Box::new(exp),
postfixes: inner.map(|p| Postfix::from_pair(p)).collect(),
postfixes: inner.map(|p| Postfix::from(p)).collect(),
}
} else {
exp
}
}
Rule::primary => Expression::from_pair(pair.into_inner().next().unwrap()),
Rule::primary => Expression::from(pair.into_inner().next().unwrap()),
Rule::identifier => Expression::Identifier(pair.as_str().to_string()),
Rule::integer => {
let value = pair.as_str().parse::<i64>().unwrap();
@@ -327,21 +337,19 @@ impl Expression {
}
}
impl Postfix {
pub fn from_pair(pair: pest::iterators::Pair<Rule>) -> Self {
impl From<pest::iterators::Pair<'_, Rule>> for Postfix {
fn from(pair: pest::iterators::Pair<Rule>) -> Self {
match pair.as_rule() {
Rule::postfix => Postfix::from_pair(pair.into_inner().next().unwrap()),
Rule::postfix => Postfix::from(pair.into_inner().next().unwrap()),
Rule::field_px => {
let field_name = pair.into_inner().next().unwrap().as_str().to_string();
Postfix::FieldAccess(field_name)
}
Rule::call_px => Postfix::Call(pair.into_inner().map(Expression::from_pair).collect()),
Rule::call_px => Postfix::Call(pair.into_inner().map(Expression::from).collect()),
Rule::index_px => {
Postfix::Index(Expression::from_pair(pair.into_inner().next().unwrap()))
}
Rule::index_px => Postfix::Index(Expression::from(pair.into_inner().next().unwrap())),
Rule::binary_px => {
let mut inner = pair.into_inner();
@@ -363,7 +371,7 @@ impl Postfix {
unimplemented!("Binary operator not implemented yet: {}", op_pair.as_str())
}
};
Postfix::Binary(op, Expression::from_pair(inner.next().unwrap()))
Postfix::Binary(op, Expression::from(inner.next().unwrap()))
}
_ => unimplemented!("Postfix parsing not implemented yet {:?}", pair.as_rule()),
@@ -373,8 +381,8 @@ impl Postfix {
fn parse_var_assign_no_semicolon(pair: pest::iterators::Pair<Rule>) -> Statement {
let mut inner = pair.into_inner();
let target = Expression::from_pair(inner.next().unwrap());
let value = Expression::from_pair(inner.next().unwrap());
let target = Expression::from(inner.next().unwrap());
let value = Expression::from(inner.next().unwrap());
Statement::VarAssign { target, value }
}
+9
View File
@@ -0,0 +1,9 @@
[package]
name = "semantic"
version = "0.1.0"
edition = "2024"
[dependencies]
parser = { path = "../parser" }
serde = { version = "1.0.228", features = ["serde_derive"] }
serde_json = "1.0.149"
+318
View File
@@ -0,0 +1,318 @@
use std::{
collections::HashMap,
sync::{
Arc,
atomic::{AtomicUsize, Ordering},
},
};
use crate::top_level::{
FunctionSymbol, JSONScope, StructSymbol, TopLevelSymbolScope, TypeSymbol, VarSymbol,
};
#[derive(Clone, Debug)]
pub enum TypeRef {
Struct(StructRef),
Function(FunctionRef),
Package(PackageRef),
Name(String),
}
#[derive(Clone, Debug)]
pub struct PackageRef {
pub name: String,
pub variables: HashMap<String, Arc<VarRef>>,
}
#[derive(Clone, Debug)]
pub struct VarRef {
pub export: bool,
pub var_type: Arc<TypeRef>,
pub name: String,
}
#[derive(Clone, Debug)]
pub struct FunctionRef {
pub export: bool,
pub name: String,
pub params: HashMap<String, Arc<VarRef>>,
pub return_type: Option<Arc<TypeRef>>,
}
#[derive(Clone, Debug)]
pub struct StructRef {
pub export: bool,
pub name: String,
pub fields: HashMap<String, Arc<VarRef>>,
}
#[derive(Debug)]
pub struct TopLevelHirScope {
pub variables: HashMap<String, Arc<VarRef>>,
pub var_idx: AtomicUsize,
}
impl TopLevelHirScope {
pub fn from_tlss(tlss: &TopLevelSymbolScope) -> Self {
let mut scope = Self {
variables: HashMap::new(),
var_idx: AtomicUsize::new(0),
};
for (_, symbol) in &tlss.functions {
scope.function_ref(tlss, symbol);
}
for (_, symbol) in &tlss.structs {
scope.struct_ref(tlss, symbol);
}
for (_, symbol) in &tlss.imports {
scope.variables.insert(
symbol.package_name.clone(),
Arc::new(VarRef {
export: false,
name: symbol.package_name.clone(),
var_type: Arc::new(TypeRef::Package(PackageRef::from_tlss(symbol))),
}),
);
}
scope
}
pub fn function_ref(&mut self, tlss: &TopLevelSymbolScope, symbol: &FunctionSymbol) {
if self.variables.get(&symbol.name).is_none() {
let name = self.get_name(symbol.export);
if let Some(_) = tlss.functions.get(&symbol.name) {
let rf = FunctionRef {
export: symbol.export,
name: name.clone(),
params: symbol
.params
.iter()
.map(|(name, v)| (name.clone(), self.var_ref(tlss, v)))
.collect(),
return_type: symbol
.return_type
.clone()
.map(|rt| self.type_ref(tlss, &rt)),
};
self.variables.insert(
symbol.name.clone(),
Arc::new(VarRef {
export: false,
name: name.clone(),
var_type: Arc::new(TypeRef::Function(rf)),
}),
);
} else {
unimplemented!()
}
}
}
pub fn struct_ref(
&mut self,
tlss: &TopLevelSymbolScope,
symbol: &StructSymbol,
) -> Arc<TypeRef> {
if let Some(rf) = self.variables.get(&symbol.name) {
rf.var_type.clone()
} else {
let name = self.get_name(symbol.export);
let rf = Arc::new(TypeRef::Struct(StructRef {
export: symbol.export,
name: name.clone(),
fields: symbol
.fields
.iter()
.map(|(name, v)| (name.clone(), self.var_ref(tlss, v)))
.collect(),
}));
self.variables.insert(
symbol.name.clone(),
Arc::new(VarRef {
export: symbol.export,
name: name.clone(),
var_type: rf.clone(),
}),
);
rf
}
}
pub fn var_ref(&mut self, tlss: &TopLevelSymbolScope, symbol: &VarSymbol) -> Arc<VarRef> {
Arc::new(VarRef {
export: symbol.export,
var_type: self.type_ref(tlss, &symbol.var_type),
name: self.get_name(symbol.export),
})
}
pub fn type_ref(&mut self, tlss: &TopLevelSymbolScope, symbol: &TypeSymbol) -> Arc<TypeRef> {
if let Some(var_ref) = self.variables.get(&symbol.0) {
var_ref.var_type.clone()
} else {
if let Some(tlss_rf) = tlss.structs.get(&symbol.0) {
self.struct_ref(tlss, tlss_rf)
} else {
let var_ref = Arc::new(VarRef {
export: false,
name: symbol.0.clone(),
var_type: Arc::new(TypeRef::Name(symbol.0.clone())),
});
self.variables.insert(symbol.0.clone(), var_ref.clone());
var_ref.var_type.clone()
}
}
}
pub fn get_reference(&self, name: &String) -> Option<Arc<VarRef>> {
self.variables.get(name).cloned()
}
pub fn next_var_idx(&self) -> usize {
self.var_idx.fetch_add(1, Ordering::Relaxed)
}
pub fn get_name(&self, export: bool) -> String {
format!("{}{}", if export { 'V' } else { 'v' }, self.next_var_idx())
}
}
impl TypeRef {
pub fn get_name(&self) -> String {
match self {
TypeRef::Function(f) => f.name.clone(),
TypeRef::Struct(s) => s.name.clone(),
TypeRef::Package(p) => p.name.clone(),
TypeRef::Name(n) => n.clone(),
}
}
}
impl PackageRef {
pub fn from_tlss(json_scope: &JSONScope) -> Self {
let mut scope = Self {
name: json_scope.package_name.clone(),
variables: HashMap::new(),
};
for (token_name, symbol) in &json_scope.functions {
scope.function_ref(json_scope, token_name, symbol);
}
for (_, symbol) in &json_scope.structs {
scope.struct_ref(json_scope, symbol);
}
scope
}
pub fn function_ref(
&mut self,
json_scope: &JSONScope,
token_name: &String,
symbol: &FunctionSymbol,
) {
if self.variables.get(&symbol.name).is_none() {
let name = symbol.name.clone();
if let Some(_) = json_scope.functions.get(token_name) {
let rf = FunctionRef {
export: symbol.export,
name: name.clone(),
params: symbol
.params
.iter()
.map(|(name, v)| (name.clone(), self.var_ref(json_scope, v)))
.collect(),
return_type: symbol
.return_type
.clone()
.map(|rt| self.type_ref(json_scope, &rt)),
};
self.variables.insert(
token_name.clone(),
Arc::new(VarRef {
export: false,
name: name.clone(),
var_type: Arc::new(TypeRef::Function(rf)),
}),
);
} else {
unimplemented!()
}
}
}
pub fn struct_ref(&mut self, json_scope: &JSONScope, symbol: &StructSymbol) -> Arc<TypeRef> {
if let Some(rf) = self.variables.get(&symbol.name) {
rf.var_type.clone()
} else {
let name = symbol.name.clone();
let rf = Arc::new(TypeRef::Struct(StructRef {
export: symbol.export,
name: name.clone(),
fields: symbol
.fields
.iter()
.map(|(name, v)| (name.clone(), self.var_ref(json_scope, v)))
.collect(),
}));
self.variables.insert(
symbol.name.clone(),
Arc::new(VarRef {
export: symbol.export,
name: name.clone(),
var_type: rf.clone(),
}),
);
rf
}
}
pub fn var_ref(&mut self, json_scope: &JSONScope, symbol: &VarSymbol) -> Arc<VarRef> {
Arc::new(VarRef {
export: symbol.export,
var_type: self.type_ref(json_scope, &symbol.var_type),
name: symbol.name.clone(),
})
}
pub fn type_ref(&mut self, json_scope: &JSONScope, symbol: &TypeSymbol) -> Arc<TypeRef> {
if let Some(var_ref) = self.variables.get(&symbol.0) {
var_ref.var_type.clone()
} else {
if let Some(json_scope_rf) = json_scope.structs.get(&symbol.0) {
self.struct_ref(json_scope, json_scope_rf)
} else {
let var_ref = Arc::new(VarRef {
export: false,
name: symbol.0.clone(),
var_type: Arc::new(TypeRef::Name(symbol.0.clone())),
});
self.variables.insert(symbol.0.clone(), var_ref.clone());
var_ref.var_type.clone()
}
}
}
pub fn get_reference(&self, name: &String) -> Option<Arc<VarRef>> {
self.variables.get(name).cloned()
}
}
+67
View File
@@ -0,0 +1,67 @@
use std::{collections::HashMap, sync::Arc};
use parser::ast::{ParamList, TypeExpr};
use crate::{
hir::VarRef,
scope::{LocalScope, Scope},
};
pub mod hir;
pub mod scope;
pub mod top_level;
pub fn walk_ast(top_scope: Arc<Scope>, tl: &mut Vec<parser::ast::TopLevel>) {
for tl in tl {
match tl {
parser::ast::TopLevel::Import(_) => {}
parser::ast::TopLevel::Package(_) => {}
parser::ast::TopLevel::FunctionDecl {
params, body, name, ..
} => {
let rf = top_scope.get_reference(name).unwrap();
if name != "main" {
*name = rf.name.clone();
}
match &*rf.var_type {
hir::TypeRef::Function(f) => walk_param_list(&f.params, params),
_ => unimplemented!(),
}
let scope = LocalScope::new(top_scope.clone());
scope.with_params(params);
scope.with_block(body);
}
parser::ast::TopLevel::StructDecl { name, fields, .. } => {
let rf = top_scope.get_reference(name).unwrap();
*name = rf.name.clone();
match &*rf.var_type {
hir::TypeRef::Struct(s) => walk_param_list(&s.fields, fields),
_ => unimplemented!(),
}
}
}
}
}
pub fn walk_param_list(fields: &HashMap<String, Arc<VarRef>>, param_list: &mut ParamList) {
let old_param_list = param_list.clone();
param_list.0.clear();
for (name, param) in fields {
param_list.0.insert(
param.name.clone(),
(
old_param_list.0.get(name).map(|a| a.0).unwrap_or_default(),
TypeExpr::Identifier(param.var_type.get_name()),
),
);
}
}
+241
View File
@@ -0,0 +1,241 @@
use std::{
collections::HashMap,
sync::{Arc, Mutex},
};
use parser::ast::{self, ParamList, Postfix, Statement};
use crate::{
hir::{TopLevelHirScope, TypeRef, VarRef},
top_level::TopLevelSymbolScope,
};
#[derive(Debug)]
pub enum Scope {
TopLevel(TopLevelHirScope),
Local(LocalScope),
}
impl Scope {
pub fn from_top(
path: &std::path::PathBuf,
top_level: &Vec<parser::ast::TopLevel>,
) -> Arc<Self> {
let tl = TopLevelSymbolScope::from(path, top_level);
Arc::new(Self::TopLevel(TopLevelHirScope::from_tlss(&tl)))
}
pub fn get_reference(&self, name: &String) -> Option<Arc<VarRef>> {
match self {
Scope::TopLevel(tl) => tl.get_reference(name),
Scope::Local(l) => l.get_reference(name),
}
}
pub fn next_var_idx(&self) -> usize {
match self {
Scope::TopLevel(tl) => tl.next_var_idx(),
Scope::Local(l) => l.parent.next_var_idx(),
}
}
pub fn get_name(&self, export: bool) -> String {
format!("{}{}", if export { 'V' } else { 'v' }, self.next_var_idx())
}
}
#[derive(Debug)]
pub struct LocalScope {
pub parent: Arc<Scope>,
pub variables: Mutex<HashMap<String, Arc<VarRef>>>,
}
impl LocalScope {
pub fn new(parent: Arc<Scope>) -> Arc<Self> {
Arc::new(LocalScope {
parent,
variables: Mutex::new(HashMap::new()),
})
}
pub fn get_reference(&self, name: &String) -> Option<Arc<VarRef>> {
let rf = self
.variables
.lock()
.unwrap()
.get(name)
.cloned()
.or_else(|| self.parent.get_reference(name));
match rf {
Some(v) => Some(v),
None => {
let var_ref = Arc::new(VarRef {
export: false,
name: name.clone(),
var_type: Arc::new(TypeRef::Name(name.clone())),
});
self.variables
.lock()
.unwrap()
.insert(name.clone(), var_ref.clone());
Some(var_ref)
}
}
}
pub fn with_statement(self: &Arc<Self>, statement: &mut Statement) {
match statement {
Statement::Block(b) => self.clone().with_block(b),
Statement::VarDecl { name, init, .. } => {
if let Some(init) = init {
let var_type = self.get_type_from_expr(init).unwrap();
let var_name = name.clone();
*name = self.parent.get_name(false);
self.variables.lock().unwrap().insert(
var_name,
Arc::new(VarRef {
export: false,
name: name.clone(),
var_type,
}),
);
}
}
Statement::Expression(e) => {
self.get_type_from_expr(e);
}
Statement::If {
condition,
then_branch,
else_branch,
} => {
self.get_type_from_expr(condition);
self.clone().with_statement(then_branch);
if let Some(else_branch) = else_branch {
self.clone().with_statement(else_branch);
}
}
_ => {}
}
}
pub fn with_block(self: &Arc<Self>, block: &mut parser::ast::Block) {
for statement in &mut block.0 {
self.with_statement(statement);
}
}
pub fn walk_postfixes(
self: &Arc<Self>,
initial: &mut Box<ast::Expression>,
postfixes: &mut Vec<Postfix>,
) -> Option<Arc<TypeRef>> {
let mut current_type = self.get_type_from_expr(initial)?;
for postfix in postfixes {
match postfix {
Postfix::FieldAccess(id) => match &*current_type {
TypeRef::Struct(s) => {
let field = s.fields.get(id)?;
*id = field.name.clone();
current_type = field.var_type.clone();
}
TypeRef::Package(p) => {
let var_ref = p.variables.get(id)?;
*id = var_ref.name.clone();
current_type = var_ref.var_type.clone();
}
_ => unimplemented!(),
},
Postfix::Call(_args) => match &*current_type {
TypeRef::Function(s) => {
// TODO: arg checking
current_type = s.return_type.clone()?;
}
_ => unimplemented!(),
},
Postfix::Binary(op, right) => match op {
parser::ast::BinaryOp::Equal
| parser::ast::BinaryOp::NotEqual
| parser::ast::BinaryOp::GreaterThan
| parser::ast::BinaryOp::LessThan
| parser::ast::BinaryOp::GreaterThanOrEqual
| parser::ast::BinaryOp::LessThanOrEqual => {
self.get_type_from_expr(right)?;
current_type = Arc::new(TypeRef::Name("bool".to_string()));
}
parser::ast::BinaryOp::Plus
| parser::ast::BinaryOp::Minus
| parser::ast::BinaryOp::Multiply
| parser::ast::BinaryOp::Divide
| parser::ast::BinaryOp::Modulo => {
self.get_type_from_expr(right)?;
}
},
_ => unimplemented!(),
}
}
Some(current_type)
}
pub fn get_type_from_expr(
self: &Arc<Self>,
expr: &mut ast::Expression,
) -> Option<Arc<TypeRef>> {
match expr {
ast::Expression::IntLiteral(_) => self
.get_reference(&"int".to_string())
.map(|r| r.var_type.clone()),
ast::Expression::FloatLiteral(_) => self
.get_reference(&"float".to_string())
.map(|r| r.var_type.clone()),
ast::Expression::BoolLiteral(_) => self
.get_reference(&"bool".to_string())
.map(|r| r.var_type.clone()),
ast::Expression::StringLiteral(_) => self
.get_reference(&"string".to_string())
.map(|r| r.var_type.clone()),
ast::Expression::Identifier(id) => {
let rf = self.get_reference(id)?;
*id = rf.name.clone();
Some(rf.var_type.clone())
}
ast::Expression::Postfix { initial, postfixes } => {
self.walk_postfixes(initial, postfixes)
}
}
}
pub fn with_params(self: &Arc<Self>, param_list: &ParamList) {
for (param_name, (export, type_expr)) in &param_list.0 {
match type_expr {
parser::ast::TypeExpr::Identifier(id) => {
if let Some(var_type) =
self.parent.get_reference(id).map(|r| r.var_type.clone())
{
self.variables.lock().unwrap().insert(
param_name.clone(),
Arc::new(VarRef {
export: *export,
var_type,
name: param_name.clone(),
}),
);
}
}
}
}
}
}
+175
View File
@@ -0,0 +1,175 @@
use std::collections::HashMap;
use std::fs;
use std::path::PathBuf;
use parser::ast::TopLevel;
use parser::ast::{ParamList, TypeExpr};
use serde::{Deserialize, Serialize};
#[derive(Clone, Debug, Serialize, Deserialize)]
pub struct TypeSymbol(pub String);
#[derive(Clone, Debug, Serialize, Deserialize)]
pub struct VarSymbol {
pub export: bool,
pub var_type: TypeSymbol,
pub name: String,
}
#[derive(Clone, Debug, Serialize, Deserialize)]
pub struct FunctionSymbol {
pub export: bool,
pub name: String,
pub params: HashMap<String, VarSymbol>,
pub return_type: Option<TypeSymbol>,
}
#[derive(Clone, Debug, Serialize, Deserialize)]
pub struct StructSymbol {
pub export: bool,
pub name: String,
pub fields: HashMap<String, VarSymbol>,
pub methods: HashMap<String, FunctionSymbol>,
}
impl TypeSymbol {
pub fn from_ast(expr: TypeExpr) -> Self {
TypeSymbol(match expr {
TypeExpr::Identifier(i) => i.to_string(),
})
}
}
impl FunctionSymbol {
pub fn from_ast(
export: bool,
name: String,
params: ParamList,
return_type: Option<TypeExpr>,
) -> Self {
Self {
export: export,
name: name.clone(),
params: params
.0
.iter()
.map(|(name, (export, v))| {
(
name.clone(),
VarSymbol {
export: *export,
name: name.clone(),
var_type: TypeSymbol(match v {
TypeExpr::Identifier(i) => i.to_string(),
}),
},
)
})
.collect(),
return_type: return_type.map(TypeSymbol::from_ast),
}
}
}
impl StructSymbol {
pub fn from_ast(export: bool, name: String, fields: ParamList) -> Self {
Self {
export,
name,
fields: fields
.0
.iter()
.map(|(name, (export, v))| {
(
name.clone(),
VarSymbol {
export: *export,
name: name.clone(),
var_type: TypeSymbol(match v {
TypeExpr::Identifier(i) => i.to_string(),
}),
},
)
})
.collect(),
// TODO - parse struct methods
methods: HashMap::new(),
}
}
}
#[derive(Clone, Debug, Serialize, Deserialize)]
pub struct JSONScope {
pub structs: HashMap<String, StructSymbol>,
pub functions: HashMap<String, FunctionSymbol>,
pub package_name: String,
}
type JsonScopeMap = HashMap<String, JSONScope>;
#[derive(Clone, Debug)]
pub struct TopLevelSymbolScope {
pub imports: HashMap<String, JSONScope>,
pub structs: HashMap<String, StructSymbol>,
pub functions: HashMap<String, FunctionSymbol>,
}
impl TopLevelSymbolScope {
pub fn from(path: &PathBuf, top_level: &Vec<TopLevel>) -> Self {
let mut scope = TopLevelSymbolScope {
imports: HashMap::new(),
structs: HashMap::new(),
functions: HashMap::new(),
};
for top in top_level {
match top {
TopLevel::Import(package) => {
let json = fs::read(path.join("mist.map.json")).unwrap();
let json_scope_map: JsonScopeMap = serde_json::from_slice(&json).unwrap();
let json_scope = json_scope_map
.get(&package.replace("\"", ""))
.unwrap()
.clone();
scope
.imports
.insert(json_scope.package_name.clone(), json_scope);
}
TopLevel::Package(_) => {}
TopLevel::FunctionDecl {
export,
name,
params,
return_type,
..
} => {
scope.functions.insert(
name.clone(),
FunctionSymbol::from_ast(
*export,
name.clone(),
params.clone(),
return_type.clone(),
),
);
}
TopLevel::StructDecl {
export,
name,
fields,
} => {
scope.structs.insert(
name.clone(),
StructSymbol::from_ast(*export, name.clone(), fields.clone()),
);
}
}
}
scope
}
}
+57 -45
View File
@@ -1,13 +1,11 @@
use std::{fs, path::PathBuf, process, time::Instant};
use crate::{gogen, parser};
use serde::Deserialize;
#[derive(Deserialize)]
struct Config {
entry: String,
out_dir: String,
src: String,
output: String,
}
pub fn build() {
@@ -23,54 +21,68 @@ pub fn build() {
// 2. load config
let config = load_config(&root);
let entry_path = root.join(&config.entry);
let out_dir = root.join(&config.out_dir);
let src = root.join(&config.src);
println!(" → entry: {}", entry_path.display());
for entry in fs::read_dir(src).unwrap() {
if let Ok(entry) = entry {
let entry_path = root.join(&entry.path());
let out_dir = root.join(&config.output);
let file_name = entry.file_name();
let file_name = file_name.to_string_lossy();
// 3. read entry file
let source = match fs::read_to_string(&entry_path) {
Ok(s) => s,
Err(e) => {
eprintln!("error: failed to read entry file\n {}", e);
process::exit(1);
let script = if file_name.ends_with(".ms") {
true
} else if file_name.ends_with(".mist") {
false
} else {
continue;
};
// 3. read entry file
let source = match fs::read_to_string(&entry_path) {
Ok(s) => s,
Err(e) => {
eprintln!("error: failed to read entry file\n {}", e);
process::exit(1);
}
};
let parser_result = if script {
parser::script_parser::parse(&source).map_err(|e| e.to_string())
} else {
parser::parse(&source).map_err(|e| e.to_string())
};
let mut ast = match parser_result {
Ok(ast) => ast,
Err(e) => {
eprintln!("error: parse failed\n{}", e);
process::exit(1);
}
};
semantic::walk_ast(semantic::scope::Scope::from_top(&root, &ast), &mut ast);
let mut gc = codegen::GoCodegen::new();
let output = gc.generate(&ast);
if let Err(e) = fs::create_dir_all(&out_dir) {
eprintln!("error: failed to create build dir\n {}", e);
process::exit(1);
}
let out_file =
out_dir.join(file_name.replace(if script { ".ms" } else { ".mist" }, ".go"));
if let Err(e) = fs::write(&out_file, output) {
eprintln!("error: failed to write output\n {}", e);
process::exit(1);
}
}
};
println!(" → parsing...");
let ast = match parser::parse(&source) {
Ok(ast) => {
println!(" ✓ parsed {} items", ast.len());
ast
}
Err(e) => {
eprintln!("error: parse failed\n{}", e);
process::exit(1);
}
};
println!(" → generating Go code...");
// let output = gogen::generate(&ast);
let output = String::from("// TODO: implement code generation\n");
// 4. ensure build dir
if let Err(e) = fs::create_dir_all(&out_dir) {
eprintln!("error: failed to create build dir\n {}", e);
process::exit(1);
}
let out_file = out_dir.join("main.go");
if let Err(e) = fs::write(&out_file, output) {
eprintln!("error: failed to write output\n {}", e);
process::exit(1);
}
let elapsed = start.elapsed();
println!(" ✓ built {}", out_file.display());
println!("build finished in {:.2?}", elapsed);
}
View File
-283
View File
@@ -1,283 +0,0 @@
// use crate::parser::ast::{
// BinOperator, Class, Expression, Function, Statement, Struct, TopLevel, TopLevel, TypeExpr,
// };
// pub fn generate(program: &TopLevel) -> String {
// let mut out = String::new();
// out.push_str("package main\n\n");
// // imports
// let mut imports = vec![];
// for item in &program.statements {
// if let TopLevel::Import(i) = item {
// imports.push(i.path.clone());
// }
// }
// if !imports.is_empty() {
// out.push_str("import (\n");
// for imp in imports {
// out.push_str(&format!(" {}\n", imp));
// }
// out.push_str(")\n\n");
// }
// // rest
// for item in &program.statements {
// match item {
// TopLevel::Function(f) => out.push_str(&gen_function(f)),
// TopLevel::Struct(s) => out.push_str(&gen_struct(s)),
// TopLevel::Class(c) => out.push_str(&gen_class(c)),
// TopLevel::Import(_) => {}
// }
// out.push('\n');
// }
// out
// }
// fn gen_struct(s: &Struct) -> String {
// let mut out = format!("type {} struct {{\n", s.name);
// for field in &s.fields {
// out.push_str(&format!(
// " {} {}\n",
// capitalize(&field.name),
// gen_type(&field.type_expr)
// ));
// }
// out.push_str("}\n");
// out
// }
// fn gen_class(c: &Class) -> String {
// let mut out = String::new();
// // struct
// out.push_str(&format!("type {} struct {{\n", c.name));
// for field in &c.fields {
// out.push_str(&format!(
// " {} {}\n",
// capitalize(&field.name),
// gen_type(&field.type_expr)
// ));
// }
// out.push_str("}\n\n");
// // methods
// for method in &c.methods {
// out.push_str(&gen_method(c, method));
// out.push('\n');
// }
// out
// }
// fn gen_function(f: &Function) -> String {
// let mut out = format!("func {}(", f.name);
// // params
// for (i, p) in f.params.iter().enumerate() {
// if i > 0 {
// out.push_str(", ");
// }
// out.push_str(&format!("{} {}", p.name, gen_type(&p.type_expr)));
// }
// out.push(')');
// // return
// if let Some(ret) = &f.return_type {
// out.push_str(&format!(" {}", gen_type(ret)));
// }
// out.push_str(" {\n");
// for stmt in &f.body {
// out.push_str(&gen_statement(stmt));
// }
// out.push_str("}\n");
// out
// }
// fn gen_method(class: &Class, f: &Function) -> String {
// let mut out = format!("func (self *{}) {}(", class.name, f.name);
// for (i, p) in f.params.iter().enumerate() {
// if i > 0 {
// out.push_str(", ");
// }
// out.push_str(&format!("{} {}", p.name, gen_type(&p.type_expr)));
// }
// out.push(')');
// if let Some(ret) = &f.return_type {
// out.push_str(&format!(" {}", gen_type(ret)));
// }
// out.push_str(" {\n");
// for stmt in &f.body {
// out.push_str(&gen_statement(stmt));
// }
// out.push_str("}\n");
// out
// }
// fn gen_statement(stmt: &Statement) -> String {
// match stmt {
// Statement::Let(s) => {
// let mut out = format!(" {} := {}", s.name, gen_expr(&s.value));
// out.push_str(";\n");
// out
// }
// Statement::Return(r) => match &r.value {
// Some(v) => format!(" return {};\n", gen_expr(v)),
// None => " return;\n".to_string(),
// },
// Statement::Expression(e) => {
// format!(" {};\n", gen_expr(e))
// }
// Statement::If(i) => {
// let mut out = format!(" if {} {{\n", gen_expr(&i.condition));
// for stmt in &i.body {
// out.push_str(&gen_statement(stmt));
// }
// out.push_str(" }");
// if let Some(else_body) = &i.else_body {
// out.push_str(" else {\n");
// for stmt in else_body {
// out.push_str(&gen_statement(stmt));
// }
// out.push_str(" }");
// }
// out.push('\n');
// out
// }
// Statement::For(f) => {
// let mut out = format!(
// " for _, {} := range {} {{\n",
// f.var,
// gen_expr(&f.iterator)
// );
// for stmt in &f.body {
// out.push_str(&gen_statement(stmt));
// }
// out.push_str(" }\n");
// out
// }
// }
// }
// fn gen_expr(expr: &Expression) -> String {
// match expr {
// Expression::Identifier(name, _) => name.clone(),
// Expression::Integer(v, _) => v.to_string(),
// Expression::Float(v, _) => v.to_string(),
// Expression::StringLit(s, _) => format!("\"{}\"", s),
// Expression::Bool(b, _) => b.to_string(),
// Expression::BinaryOp(b) => format!(
// "{} {} {}",
// gen_expr(&b.left),
// op_to_str(&b.op),
// gen_expr(&b.right)
// ),
// Expression::Call(c) => {
// let args = c.args.iter().map(gen_expr).collect::<Vec<_>>().join(", ");
// format!("{}({})", gen_expr(&c.callee), args)
// }
// Expression::FieldAccess(f) => {
// format!("{}.{}", gen_expr(&f.object), capitalize(&f.field))
// }
// Expression::StructInit(s) => {
// let mut out = format!("{}{{", s.name);
// for (i, (name, val)) in s.fields.iter().enumerate() {
// if i > 0 {
// out.push_str(", ");
// }
// out.push_str(&format!("{}: {}", capitalize(name), gen_expr(val)));
// }
// out.push('}');
// out
// }
// Expression::ArrayLiteral(arr) => {
// let elems = arr
// .elements
// .iter()
// .map(gen_expr)
// .collect::<Vec<_>>()
// .join(", ");
// format!("[]any{{{}}}", elems) // simple version
// }
// _ => todo!(),
// }
// }
// fn gen_type(t: &TypeExpr) -> String {
// match t {
// TypeExpr::Named(n) => match n.as_str() {
// "int" => "int".into(),
// "float" => "float64".into(),
// "string" => "string".into(),
// _ => n.clone(),
// },
// TypeExpr::Array(inner) => {
// format!("[]{}", gen_type(inner))
// }
// TypeExpr::Optional(inner) => {
// format!("*{}", gen_type(inner)) // pointer for optional
// }
// }
// }
// fn capitalize(s: &str) -> String {
// let mut chars = s.chars();
// match chars.next() {
// Some(c) => c.to_uppercase().collect::<String>() + chars.as_str(),
// None => String::new(),
// }
// }
// fn op_to_str(op: &BinOperator) -> &'static str {
// match op {
// BinOperator::Add => "+",
// BinOperator::Sub => "-",
// BinOperator::Mul => "*",
// BinOperator::Div => "/",
// BinOperator::Eq => "==",
// BinOperator::NotEq => "!=",
// BinOperator::Lt => "<",
// BinOperator::Gt => ">",
// BinOperator::LtEq => "<=",
// BinOperator::GtEq => ">=",
// BinOperator::And => "&&",
// BinOperator::Or => "||",
// }
// }
View File
+3 -78
View File
@@ -1,12 +1,7 @@
pub mod compiler;
pub mod errors;
pub mod gogen;
pub mod parser;
use std::fs;
use std::path::PathBuf;
use std::process;
pub mod compiler;
fn main() {
let args: Vec<String> = std::env::args().collect();
@@ -21,21 +16,7 @@ fn main() {
eprintln!("error: expected a file path\n usage: mist build");
process::exit(1);
}
cmd_build();
}
"check" => {
if args.len() < 3 {
eprintln!("error: expected a file path\n usage: mist check <file.ms>");
process::exit(1);
}
cmd_check(&args[2]);
}
"parse" => {
if args.len() < 3 {
eprintln!("error: expected a file path\n usage: mist check <file.ms>");
process::exit(1);
}
cmd_parse(&args[2]);
compiler::build();
}
"version" | "--version" | "-v" => {
println!("mist {}", env!("CARGO_PKG_VERSION"));
@@ -51,62 +32,6 @@ fn main() {
}
}
pub fn cmd_build() {
compiler::build();
}
fn cmd_check(path: &str) {
let source = read_ms_file(path);
match parser::parse(&source) {
Ok(_) => {
println!("ok");
}
Err(e) => {
eprintln!("parse error:\n{}", e);
process::exit(1);
}
}
}
fn cmd_parse(path: &str) {
let source = read_ms_file(path);
match parser::parse(&source) {
Ok(output) => {
fs::write(
"output.json",
serde_json::to_string_pretty(&output).unwrap(),
)
.unwrap_or_else(|e| {
eprintln!("error: could not write output.json: {}", e);
process::exit(1);
});
}
Err(e) => {
eprintln!("parse error:\n{}", e);
process::exit(1);
}
}
}
fn read_ms_file(path: &str) -> String {
let pb = PathBuf::from(path);
if !pb.exists() {
eprintln!("error: file '{}' not found", path);
process::exit(1);
}
if pb.extension().and_then(|e| e.to_str()) != Some("ms") {
eprintln!("error: expected a .ms file, got '{}'", path);
process::exit(1);
}
fs::read_to_string(&pb).unwrap_or_else(|e| {
eprintln!("error: could not read '{}': {}", path, e);
process::exit(1);
})
}
fn print_usage() {
println!("mist - the mist compiler");
println!();
+2 -2
View File
@@ -1,4 +1,4 @@
{
"entry": "src/main.ms",
"out_dir": "build"
"src": "src",
"output": "build"
}
+450
View File
@@ -0,0 +1,450 @@
{
"fmt": {
"package_name": "fmt",
"structs": {},
"functions": {
"append": {
"export": true,
"name": "Append",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
},
"b": {
"export": true,
"var_type": "[]byte",
"name": "b"
}
},
"return_types": [
"[]byte"
]
},
"appendf": {
"export": true,
"name": "Appendf",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
},
"b": {
"export": true,
"var_type": "[]byte",
"name": "b"
},
"format": {
"export": true,
"var_type": "string",
"name": "format"
}
},
"return_types": [
"[]byte"
]
},
"appendln": {
"export": true,
"name": "Appendln",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
},
"b": {
"export": true,
"var_type": "[]byte",
"name": "b"
}
},
"return_types": [
"[]byte"
]
},
"errorf": {
"export": true,
"name": "Errorf",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
},
"format": {
"export": true,
"var_type": "string",
"name": "format"
}
},
"return_types": [
"error"
]
},
"formatString": {
"export": true,
"name": "FormatString",
"params": {
"state": {
"export": true,
"var_type": "fmt.State",
"name": "state"
},
"verb": {
"export": true,
"var_type": "rune",
"name": "verb"
}
},
"return_types": [
"string"
]
},
"fprint": {
"export": true,
"name": "Fprint",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
},
"w": {
"export": true,
"var_type": "io.Writer",
"name": "w"
}
},
"return_types": [
"int",
"error"
]
},
"fprintf": {
"export": true,
"name": "Fprintf",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
},
"format": {
"export": true,
"var_type": "string",
"name": "format"
},
"w": {
"export": true,
"var_type": "io.Writer",
"name": "w"
}
},
"return_types": [
"int",
"error"
]
},
"fprintln": {
"export": true,
"name": "Fprintln",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
},
"w": {
"export": true,
"var_type": "io.Writer",
"name": "w"
}
},
"return_types": [
"int",
"error"
]
},
"fscan": {
"export": true,
"name": "Fscan",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
},
"r": {
"export": true,
"var_type": "io.Reader",
"name": "r"
}
},
"return_types": [
"int",
"error"
]
},
"fscanf": {
"export": true,
"name": "Fscanf",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
},
"format": {
"export": true,
"var_type": "string",
"name": "format"
},
"r": {
"export": true,
"var_type": "io.Reader",
"name": "r"
}
},
"return_types": [
"int",
"error"
]
},
"fscanln": {
"export": true,
"name": "Fscanln",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
},
"r": {
"export": true,
"var_type": "io.Reader",
"name": "r"
}
},
"return_types": [
"int",
"error"
]
},
"print": {
"export": true,
"name": "Print",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
}
},
"return_types": [
"int",
"error"
]
},
"printf": {
"export": true,
"name": "Printf",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
},
"format": {
"export": true,
"var_type": "string",
"name": "format"
}
},
"return_types": [
"int",
"error"
]
},
"println": {
"export": true,
"name": "Println",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
}
},
"return_types": [
"int",
"error"
]
},
"scan": {
"export": true,
"name": "Scan",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
}
},
"return_types": [
"int",
"error"
]
},
"scanf": {
"export": true,
"name": "Scanf",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
},
"format": {
"export": true,
"var_type": "string",
"name": "format"
}
},
"return_types": [
"int",
"error"
]
},
"scanln": {
"export": true,
"name": "Scanln",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
}
},
"return_types": [
"int",
"error"
]
},
"sprint": {
"export": true,
"name": "Sprint",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
}
},
"return_types": [
"string"
]
},
"sprintf": {
"export": true,
"name": "Sprintf",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
},
"format": {
"export": true,
"var_type": "string",
"name": "format"
}
},
"return_types": [
"string"
]
},
"sprintln": {
"export": true,
"name": "Sprintln",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
}
},
"return_types": [
"string"
]
},
"sscan": {
"export": true,
"name": "Sscan",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
},
"str": {
"export": true,
"var_type": "string",
"name": "str"
}
},
"return_types": [
"int",
"error"
]
},
"sscanf": {
"export": true,
"name": "Sscanf",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
},
"format": {
"export": true,
"var_type": "string",
"name": "format"
},
"str": {
"export": true,
"var_type": "string",
"name": "str"
}
},
"return_types": [
"int",
"error"
]
},
"sscanln": {
"export": true,
"name": "Sscanln",
"params": {
"a": {
"export": true,
"var_type": "[]any",
"name": "a"
},
"str": {
"export": true,
"var_type": "string",
"name": "str"
}
},
"return_types": [
"int",
"error"
]
}
}
}
}
+8
View File
@@ -0,0 +1,8 @@
package main;
import "fmt";
void main() {
string test = "Hello, World";
fmt.println(test);
}
-17
View File
@@ -1,17 +0,0 @@
function add(a: int, b: int): int {
let sum = 0;
for (let i = 0; i < 10; i = i + 1) {
if (i == 5) {
continue;
}
sum = sum + i;
}
if (sum > 20) {
return sum;
} else {
return 0;
}
}