语法/前端解析,构建AST

This commit is contained in:
jing
2025-12-29 16:24:29 +08:00
parent 8903bf73f9
commit c153604c2e
8 changed files with 388 additions and 36 deletions

View File

@@ -1,5 +1,34 @@
// 前端解析驱动:
// - 读取源代码
// - 调用 ANTLR 生成的 lexer/parser 得到 parse tree
// - 对外提供“可用的解析入口”(语法正确性由测试保证)
// 调用 ANTLR 生成的 Lexer/Parser返回 parse tree。
#include "frontend/AntlrDriver.h"
#include <fstream>
#include <sstream>
#include <stdexcept>
#include "SysYLexer.h"
#include "SysYParser.h"
#include "antlr4-runtime.h"
AntlrResult ParseFileWithAntlr(const std::string& path) {
std::ifstream fin(path);
if (!fin.is_open()) {
throw std::runtime_error("无法打开输入文件: " + path);
}
std::ostringstream ss;
ss << fin.rdbuf();
auto input = std::make_unique<antlr4::ANTLRInputStream>(ss.str());
auto lexer = std::make_unique<SysYLexer>(input.get());
auto tokens = std::make_unique<antlr4::CommonTokenStream>(lexer.get());
auto parser = std::make_unique<SysYParser>(tokens.get());
parser->removeErrorListeners();
auto tree = parser->compUnit();
AntlrResult result;
result.input = std::move(input);
result.lexer = std::move(lexer);
result.tokens = std::move(tokens);
result.parser = std::move(parser);
result.tree = tree;
return result;
}

View File

@@ -0,0 +1,20 @@
// 包装 ANTLR4提供简易的解析入口。
#pragma once
#include <memory>
#include <string>
#include "SysYLexer.h"
#include "SysYParser.h"
#include "antlr4-runtime.h"
struct AntlrResult {
std::unique_ptr<antlr4::ANTLRInputStream> input;
std::unique_ptr<SysYLexer> lexer;
std::unique_ptr<antlr4::CommonTokenStream> tokens;
std::unique_ptr<SysYParser> parser;
antlr4::tree::ParseTree* tree = nullptr; // owned by parser
};
// 解析指定文件,发生错误时抛出 std::runtime_error。
AntlrResult ParseFileWithAntlr(const std::string& path);

View File

@@ -1,4 +1,114 @@
// AST 构建:
// - 将 ANTLR parse tree 转换为 AST对应 src/ast/*
// - 在 AST 节点上保留必要的定位信息(可选,用于调试/日志)
// 将 parse tree 转换为 AST。
#include "frontend/AstBuilder.h"
#include <any>
#include <stdexcept>
#include <string>
#include <utility>
#include <vector>
#include "SysYBaseVisitor.h"
#include "SysYParser.h"
#include "ast/AstNodes.h"
#include "antlr4-runtime.h"
namespace {
using ast::BinaryExpr;
using ast::BinaryOp;
using ast::Block;
using ast::CompUnit;
using ast::FuncDef;
using ast::NumberExpr;
using ast::ReturnStmt;
using ast::VarDecl;
using ast::VarExpr;
template <typename T>
T Take(std::any&& value) {
if (auto* ptr = std::any_cast<T>(&value)) {
return std::move(*ptr);
}
throw std::runtime_error("AST 构建失败:类型不匹配");
}
class Builder : public SysYBaseVisitor {
public:
std::any visitCompUnit(SysYParser::CompUnitContext* ctx) override {
auto func = Take<std::shared_ptr<FuncDef>>(visit(ctx->funcDef()));
return std::make_shared<CompUnit>(std::move(func));
}
std::any visitFuncDef(SysYParser::FuncDefContext* ctx) override {
auto body = Take<std::shared_ptr<Block>>(visit(ctx->block()));
return std::make_shared<FuncDef>("main", std::move(body));
}
std::any visitBlock(SysYParser::BlockContext* ctx) override {
auto block = std::make_shared<Block>();
for (auto stmtCtx : ctx->stmt()) {
if (stmtCtx->varDecl()) {
block->varDecls.emplace_back(
Take<std::shared_ptr<VarDecl>>(visit(stmtCtx->varDecl())));
} else if (stmtCtx->returnStmt()) {
block->stmts.emplace_back(
Take<std::shared_ptr<ReturnStmt>>(visit(stmtCtx->returnStmt())));
}
}
return block;
}
std::any visitVarDecl(SysYParser::VarDeclContext* ctx) override {
std::shared_ptr<ast::Expr> init;
if (ctx->exp()) {
init = Take<std::shared_ptr<ast::Expr>>(visit(ctx->exp()));
}
return std::make_shared<VarDecl>(ctx->Ident()->getText(), std::move(init));
}
std::any visitReturnStmt(SysYParser::ReturnStmtContext* ctx) override {
auto expr = Take<std::shared_ptr<ast::Expr>>(visit(ctx->exp()));
return std::make_shared<ReturnStmt>(std::move(expr));
}
std::any visitExp(SysYParser::ExpContext* ctx) override {
return visit(ctx->addExp());
}
std::any visitAddExp(SysYParser::AddExpContext* ctx) override {
auto node = Take<std::shared_ptr<ast::Expr>>(visit(ctx->primary(0)));
for (size_t i = 1; i < ctx->primary().size(); ++i) {
auto rhs = Take<std::shared_ptr<ast::Expr>>(visit(ctx->primary(i)));
auto opToken = ctx->AddOp(i - 1);
BinaryOp op = BinaryOp::Add;
if (opToken->getText() == "-") op = BinaryOp::Sub;
node = std::make_shared<BinaryExpr>(op, std::move(node), std::move(rhs));
}
return node;
}
std::any visitPrimary(SysYParser::PrimaryContext* ctx) override {
if (ctx->Number()) {
std::shared_ptr<ast::Expr> expr =
std::make_shared<NumberExpr>(std::stoi(ctx->Number()->getText()));
return expr;
}
if (ctx->Ident()) {
std::shared_ptr<ast::Expr> expr =
std::make_shared<VarExpr>(ctx->Ident()->getText());
return expr;
}
return visit(ctx->exp());
}
};
} // namespace
std::shared_ptr<ast::CompUnit> BuildAst(antlr4::tree::ParseTree* tree) {
if (!tree) {
throw std::runtime_error("parse tree 为空");
}
Builder visitor;
auto result = visitor.visit(tree);
return Take<std::shared_ptr<ast::CompUnit>>(std::move(result));
}

16
src/frontend/AstBuilder.h Normal file
View File

@@ -0,0 +1,16 @@
// 将 ANTLR parse tree 转换为内部 AST。
#pragma once
#include <memory>
namespace antlr4 {
namespace tree {
class ParseTree;
}
} // namespace antlr4
namespace ast {
struct CompUnit;
}
std::shared_ptr<ast::CompUnit> BuildAst(antlr4::tree::ParseTree* tree);