CompileThreeMaggot/nudt-compiler-cpp
3
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge branch 'master' of https://gitee.com/NUDT-compiler/nudt-compiler-cpp

Browse Source
This commit is contained in:
jing
2026-03-12 00:57:57 +08:00
parent 2737ebd9fc e9d7b4a058
commit 7f1b0aaead
26 changed files with 156 additions and 339 deletions
Download Patch File Download Diff File Expand all files Collapse all files

25
src/frontend/AntlrDriver.cpp
View File

@@ -9,20 +9,17 @@
#include "SysYLexer.h"
#include "SysYParser.h"
#include "antlr4-runtime.h"
#include "utils/Log.h"
namespace {
bool HasParsePrefix(const std::string& msg) {
return msg.rfind("[parse]", 0) == 0;
}
class ParseErrorListener : public antlr4::BaseErrorListener {
public:
void syntaxError(antlr4::Recognizer* /*recognizer*/, antlr4::Token* /*offendingSymbol*/,
size_t line, size_t charPositionInLine,
const std::string& msg, std::exception_ptr /*e*/) override {
throw std::runtime_error("[parse] 暂不支持的语法/词法 @" + std::to_string(line) + ":" +
std::to_string(charPositionInLine) + " - " + msg);
throw std::runtime_error(FormatErrorAt("parse", line, charPositionInLine,
"暂不支持的语法/词法 - " + msg));
}
};
@@ -31,7 +28,7 @@ class ParseErrorListener : public antlr4::BaseErrorListener {
AntlrResult ParseFileWithAntlr(const std::string& path) {
std::ifstream fin(path);
if (!fin.is_open()) {
throw std::runtime_error("[parse] 无法打开输入文件: " + path);
throw std::runtime_error(FormatError("parse", "无法打开输入文件: " + path));
}
std::ostringstream ss;
ss << fin.rdbuf();
@@ -53,18 +50,18 @@ AntlrResult ParseFileWithAntlr(const std::string& path) {
} catch (const std::exception& ex) {
const std::string msg = ex.what();
if (!msg.empty()) {
if (HasParsePrefix(msg)) {
if (HasErrorPrefix(msg, "parse")) {
throw;
}
throw std::runtime_error("[parse] 暂不支持的语法/词法 - " + msg);
throw std::runtime_error(
FormatError("parse", "暂不支持的语法/词法 - " + msg));
}
if (auto* tok = parser->getCurrentToken()) {
throw std::runtime_error("[parse] 暂不支持的语法/词法 @" +
std::to_string(tok->getLine()) + ":" +
std::to_string(tok->getCharPositionInLine()) +
" near token '" + tok->getText() + "'");
throw std::runtime_error(
FormatErrorAt("parse", tok->getLine(), tok->getCharPositionInLine(),
"暂不支持的语法/词法 near token '" + tok->getText() + "'"));
}
throw std::runtime_error("[parse] 暂不支持的语法/词法");
throw std::runtime_error(FormatError("parse", "暂不支持的语法/词法"));
}
AntlrResult result;

7
src/irgen/IRGenDecl.cpp
View File

@@ -4,6 +4,7 @@
#include "SysYParser.h"
#include "ir/IR.h"
#include "utils/Log.h"
void IRGenImpl::GenBlock(SysYParser::BlockContext& block) {
for (auto* item : block.blockItem()) {
@@ -24,7 +25,7 @@ bool IRGenImpl::GenBlockItem(SysYParser::BlockItemContext& item) {
if (item.stmt()) {
return GenStmt(*item.stmt());
}
throw std::runtime_error("[irgen] 暂不支持的 blockItem 类型");
throw std::runtime_error(FormatError("irgen", "暂不支持的语句或声明"));
}
void IRGenImpl::GenDecl(SysYParser::DeclContext& decl) {
@@ -32,12 +33,12 @@ void IRGenImpl::GenDecl(SysYParser::DeclContext& decl) {
GenVarDecl(*decl.varDecl());
return;
}
throw std::runtime_error("[irgen] 暂不支持的声明类型");
throw std::runtime_error(FormatError("irgen", "暂不支持的声明类型"));
}
void IRGenImpl::GenVarDecl(SysYParser::VarDeclContext& decl) {
if (storage_map_.find(&decl) != storage_map_.end()) {
throw std::runtime_error("[irgen] 声明重复生成存储槽位");
throw std::runtime_error(FormatError("irgen", "声明重复生成存储槽位"));
}
auto* slot = builder_.CreateAllocaI32(module_.context().NextTemp());
storage_map_[&decl] = slot;

2
src/irgen/IRGenDriver.cpp
View File

@@ -1,10 +1,10 @@
#include "irgen/IRGen.h"
#include <memory>
#include <stdexcept>
#include "SysYParser.h"
#include "ir/IR.h"
#include "utils/Log.h"
std::unique_ptr<ir::Module> GenerateIR(SysYParser::CompUnitContext& tree,
const SemanticContext& sema) {

17
src/irgen/IRGenExp.cpp
View File

@@ -4,10 +4,11 @@
#include "SysYParser.h"
#include "ir/IR.h"
#include "utils/Log.h"
ir::Value* IRGenImpl::GenExpr(SysYParser::ExpContext& expr) {
if (!expr.addExp()) {
throw std::runtime_error("[irgen] 非法表达式");
throw std::runtime_error(FormatError("irgen", "非法表达式"));
}
return GenAddExpr(*expr.addExp());
}
@@ -16,7 +17,7 @@ ir::Value* IRGenImpl::GenAddExpr(SysYParser::AddExpContext& add) {
// 当前表达式层次仍是最小实现，直接贴合 addExp -> primary 的语法形状。
const auto& terms = add.primary();
if (terms.empty()) {
throw std::runtime_error("[irgen] 空加法表达式");
throw std::runtime_error(FormatError("irgen", "空加法表达式"));
}
ir::Value* acc = GenPrimary(*terms[0]);
@@ -35,18 +36,20 @@ ir::Value* IRGenImpl::GenPrimary(SysYParser::PrimaryContext& primary) {
if (primary.Ident()) {
auto* decl = sema_.ResolveVarUse(&primary);
if (!decl) {
throw std::runtime_error("[irgen] 变量使用缺少语义绑定: " +
primary.Ident()->getText());
throw std::runtime_error(
FormatError("irgen",
"变量使用缺少语义绑定: " + primary.Ident()->getText()));
}
auto it = storage_map_.find(decl);
if (it == storage_map_.end()) {
throw std::runtime_error("[irgen] 变量声明缺少存储槽位: " +
primary.Ident()->getText());
throw std::runtime_error(
FormatError("irgen",
"变量声明缺少存储槽位: " + primary.Ident()->getText()));
}
return builder_.CreateLoad(it->second, module_.context().NextTemp());
}
if (primary.exp()) {
return GenExpr(*primary.exp());
}
throw std::runtime_error("[irgen] 暂不支持的 primary 形式");
throw std::runtime_error(FormatError("irgen", "暂不支持的表达式形式"));
}

12
src/irgen/IRGenFunc.cpp
View File

@@ -4,6 +4,7 @@
#include "SysYParser.h"
#include "ir/IR.h"
#include "utils/Log.h"
namespace {
@@ -11,8 +12,9 @@ void VerifyFunctionStructure(const ir::Function& func) {
// 当前 IRGen 仍是单入口、顺序生成；这里在生成结束后补一层块终结校验。
for (const auto& bb : func.blocks()) {
if (!bb || !bb->HasTerminator()) {
throw std::runtime_error("[irgen] 基本块未正确终结: " +
(bb ? bb->name() : std::string("<null>")));
throw std::runtime_error(
FormatError("irgen", "基本块未正确终结: " +
(bb ? bb->name() : std::string("<null>"))));
}
}
}
@@ -27,17 +29,17 @@ IRGenImpl::IRGenImpl(ir::Module& module, const SemanticContext& sema)
void IRGenImpl::Gen(SysYParser::CompUnitContext& cu) {
if (!cu.funcDef()) {
throw std::runtime_error("[irgen] 缺少 main 定义");
throw std::runtime_error(FormatError("irgen", "缺少 main 定义"));
}
GenFuncDef(*cu.funcDef());
}
void IRGenImpl::GenFuncDef(SysYParser::FuncDefContext& func) {
if (!func.block()) {
throw std::runtime_error("[irgen] 函数体为空");
throw std::runtime_error(FormatError("irgen", "函数体为空"));
}
if (!func.Ident()) {
throw std::runtime_error("[irgen] 缺少函数名");
throw std::runtime_error(FormatError("irgen", "缺少函数名"));
}
func_ = module_.CreateFunction(

5
src/irgen/IRGenStmt.cpp
View File

@@ -4,18 +4,19 @@
#include "SysYParser.h"
#include "ir/IR.h"
#include "utils/Log.h"
bool IRGenImpl::GenStmt(SysYParser::StmtContext& stmt) {
if (stmt.returnStmt()) {
GenReturnStmt(*stmt.returnStmt());
return true;
}
throw std::runtime_error("[irgen] 暂不支持的语句类型");
throw std::runtime_error(FormatError("irgen", "暂不支持的语句类型"));
}
void IRGenImpl::GenReturnStmt(SysYParser::ReturnStmtContext& ret) {
if (!ret.exp()) {
throw std::runtime_error("[irgen] return 缺少表达式");
throw std::runtime_error(FormatError("irgen", "return 缺少表达式"));
}
ir::Value* v = GenExpr(*ret.exp());
builder_.CreateRet(v);

4
src/main.cpp
View File

@@ -28,7 +28,7 @@ int main(int argc, char** argv) {
auto* comp_unit = dynamic_cast<SysYParser::CompUnitContext*>(antlr.tree);
if (!comp_unit) {
throw std::runtime_error("[main] 语法树根节点不是 compUnit");
throw std::runtime_error(FormatError("main", "语法树根节点不是 compUnit"));
}
auto sema = RunSema(*comp_unit);
@@ -52,7 +52,7 @@ int main(int argc, char** argv) {
mir::PrintAsm(*machine_func, std::cout);
}
} catch (const std::exception& ex) {
std::cerr << "error: " << ex.what() << "\n";
PrintException(std::cerr, ex);
return 1;
}
return 0;

5
src/mir/FrameLowering.cpp
View File

@@ -3,6 +3,8 @@
#include <stdexcept>
#include <vector>
#include "utils/Log.h"
namespace mir {
namespace {
@@ -17,8 +19,7 @@ void RunFrameLowering(MachineFunction& function) {
for (const auto& slot : function.frame_slots()) {
cursor += slot.size;
if (-cursor < -256) {
throw std::runtime_error(
"Lab3 MVP 后端暂不支持超过 64 个 i32 栈槽的函数");
throw std::runtime_error(FormatError("mir", "暂不支持过大的栈帧"));
}
}

21
src/mir/Lowering.cpp
View File

@@ -4,6 +4,7 @@
#include <unordered_map>
#include "ir/IR.h"
#include "utils/Log.h"
namespace mir {
namespace {
@@ -20,8 +21,8 @@ void EmitValueToReg(const ir::Value* value, PhysReg target,
auto it = slots.find(value);
if (it == slots.end()) {
throw std::runtime_error("Lab3 MVP 后端找不到值对应的栈槽: " +
value->name());
throw std::runtime_error(
FormatError("mir", "找不到值对应的栈槽: " + value->name()));
}
block.Append(Opcode::LoadStack,
@@ -41,7 +42,8 @@ void LowerInstruction(const ir::Instruction& inst, MachineFunction& function,
auto& store = static_cast<const ir::StoreInst&>(inst);
auto dst = slots.find(store.ptr());
if (dst == slots.end()) {
throw std::runtime_error("Lab3 MVP 后端要求 store 目标必须来自 alloca");
throw std::runtime_error(
FormatError("mir", "暂不支持对非栈变量地址进行写入"));
}
EmitValueToReg(store.value(), PhysReg::W8, slots, block);
block.Append(Opcode::StoreStack,
@@ -52,7 +54,8 @@ void LowerInstruction(const ir::Instruction& inst, MachineFunction& function,
auto& load = static_cast<const ir::LoadInst&>(inst);
auto src = slots.find(load.ptr());
if (src == slots.end()) {
throw std::runtime_error("Lab3 MVP 后端要求 load 源必须来自 alloca");
throw std::runtime_error(
FormatError("mir", "暂不支持对非栈变量地址进行读取"));
}
int dst_slot = function.CreateFrameIndex();
block.Append(Opcode::LoadStack,
@@ -83,10 +86,10 @@ void LowerInstruction(const ir::Instruction& inst, MachineFunction& function,
}
case ir::Opcode::Sub:
case ir::Opcode::Mul:
throw std::runtime_error("Lab3 MVP 后端暂不支持 add 以外的二元运算");
throw std::runtime_error(FormatError("mir", "暂不支持该二元运算"));
}
throw std::runtime_error("Lab3 MVP 后端遇到未知 IR 指令");
throw std::runtime_error(FormatError("mir", "暂不支持该 IR 指令"));
}
} // namespace
@@ -95,19 +98,19 @@ std::unique_ptr<MachineFunction> LowerToMIR(const ir::Module& module) {
DefaultContext();
if (module.functions().size() != 1) {
throw std::runtime_error("Lab3 MVP 后端只支持单函数 main");
throw std::runtime_error(FormatError("mir", "暂不支持多个函数"));
}
const auto& func = *module.functions().front();
if (func.name() != "main") {
throw std::runtime_error("Lab3 MVP 后端只支持 main 函数");
throw std::runtime_error(FormatError("mir", "暂不支持非 main 函数"));
}
auto machine_func = std::make_unique<MachineFunction>(func.name());
ValueSlotMap slots;
const auto* entry = func.entry();
if (!entry) {
throw std::runtime_error("IR 函数缺少入口基本块");
throw std::runtime_error(FormatError("mir", "IR 函数缺少入口基本块"));
}
for (const auto& inst : entry->instructions()) {

4
src/mir/RegAlloc.cpp
View File

@@ -2,6 +2,8 @@
#include <stdexcept>
#include "utils/Log.h"
namespace mir {
namespace {
@@ -24,7 +26,7 @@ void RunRegAlloc(MachineFunction& function) {
for (const auto& inst : function.entry().instructions()) {
for (const auto& operand : inst.operands()) {
if (operand.kind() == Operand::Kind::Reg && !IsAllowedReg(operand.reg())) {
throw std::runtime_error("Lab3 MVP 后端发现未预着色的寄存器");
throw std::runtime_error(FormatError("mir", "寄存器分配失败"));
}
}
}

26
src/sem/Sema.cpp
View File

@@ -4,6 +4,7 @@
#include <string>
#include "sem/SymbolTable.h"
#include "utils/Log.h"
namespace {
@@ -20,7 +21,7 @@ void CheckPrimary(SysYParser::PrimaryContext& primary,
const std::string name = primary.Ident()->getText();
auto* decl = table.Lookup(name);
if (!decl) {
throw std::runtime_error("[sema] 使用了未定义的变量: " + name);
throw std::runtime_error(FormatError("sema", "使用了未定义的变量: " + name));
}
sema.BindVarUse(&primary, decl);
return;
@@ -31,13 +32,13 @@ void CheckPrimary(SysYParser::PrimaryContext& primary,
return;
}
throw std::runtime_error("[sema] 暂不支持的 primary 形式");
throw std::runtime_error(FormatError("sema", "暂不支持的表达式形式"));
}
void CheckExpr(SysYParser::ExpContext& exp, const SymbolTable& table,
SemanticContext& sema) {
if (!exp.addExp()) {
throw std::runtime_error("[sema] 非法表达式");
throw std::runtime_error(FormatError("sema", "非法表达式"));
}
const auto& terms = exp.addExp()->primary();
for (auto* term : terms) {
@@ -50,10 +51,10 @@ void CheckExpr(SysYParser::ExpContext& exp, const SymbolTable& table,
SemanticContext RunSema(SysYParser::CompUnitContext& comp_unit) {
auto* func = comp_unit.funcDef();
if (!func || !func->block()) {
throw std::runtime_error("[sema] 缺少 main 函数定义");
throw std::runtime_error(FormatError("sema", "缺少 main 函数定义"));
}
if (!func->Ident() || func->Ident()->getText() != "main") {
throw std::runtime_error("[sema] 入口函数必须命名为 main");
throw std::runtime_error(FormatError("sema", "入口函数必须命名为 main"));
}
SymbolTable table;
@@ -62,7 +63,8 @@ SemanticContext RunSema(SysYParser::CompUnitContext& comp_unit) {
const auto& items = func->block()->blockItem();
if (items.empty()) {
throw std::runtime_error("[sema] main 函数不能为空，且必须以 return 结束");
throw std::runtime_error(
FormatError("sema", "main 函数不能为空，且必须以 return 结束"));
}
for (size_t i = 0; i < items.size(); ++i) {
@@ -71,12 +73,13 @@ SemanticContext RunSema(SysYParser::CompUnitContext& comp_unit) {
continue;
}
if (seen_return) {
throw std::runtime_error("[sema] return 必须是 main 函数中的最后一条语句");
throw std::runtime_error(
FormatError("sema", "return 必须是 main 函数中的最后一条语句"));
}
if (auto* decl = item->decl() ? item->decl()->varDecl() : nullptr) {
const std::string name = decl->Ident()->getText();
if (table.Contains(name)) {
throw std::runtime_error("[sema] 重复定义变量: " + name);
throw std::runtime_error(FormatError("sema", "重复定义变量: " + name));
}
if (decl->exp()) {
CheckExpr(*decl->exp(), table, sema);
@@ -89,15 +92,16 @@ SemanticContext RunSema(SysYParser::CompUnitContext& comp_unit) {
CheckExpr(*ret->exp(), table, sema);
seen_return = true;
if (i + 1 != items.size()) {
throw std::runtime_error("[sema] return 必须是 main 函数中的最后一条语句");
throw std::runtime_error(
FormatError("sema", "return 必须是 main 函数中的最后一条语句"));
}
continue;
}
throw std::runtime_error("[sema] 暂不支持的 blockItem 类型");
throw std::runtime_error(FormatError("sema", "暂不支持的语句或声明"));
}
if (!seen_return) {
throw std::runtime_error("[sema] main 函数必须包含 return 语句");
throw std::runtime_error(FormatError("sema", "main 函数必须包含 return 语句"));
}
return sema;

25
src/utils/CLI.cpp
View File

@@ -5,17 +5,20 @@
#include "utils/CLI.h"
#include <cstring>
#include <stdexcept>
#include <string>
#include <cstring>
#include "utils/Log.h"
CLIOptions ParseCLI(int argc, char** argv) {
CLIOptions opt;
bool explicit_emit = false;
if (argc <= 1) {
throw std::runtime_error(
"用法: compiler [--help] [--emit-parse-tree] [--emit-ir] [--emit-asm] <input.sy>");
throw std::runtime_error(FormatError(
"cli",
"用法: compiler [--help] [--emit-parse-tree] [--emit-ir] [--emit-asm] <input.sy>"));
}
for (int i = 1; i < argc; ++i) {
@@ -59,23 +62,25 @@ CLIOptions ParseCLI(int argc, char** argv) {
}
if (arg[0] == '-') {
throw std::runtime_error(std::string("未知参数: ") + arg +
"（使用 --help 查看用法）");
throw std::runtime_error(
FormatError("cli", std::string("未知参数: ") + arg +
"（使用 --help 查看用法）"));
}
if (!opt.input.empty()) {
throw std::runtime_error(
"参数过多：当前只支持 1 个输入文件（使用 --help 查看用法）");
throw std::runtime_error(FormatError(
"cli", "参数过多：当前只支持 1 个输入文件（使用 --help 查看用法）"));
}
opt.input = arg;
}
if (opt.input.empty() && !opt.show_help) {
throw std::runtime_error("缺少输入文件：请提供 <input.sy>（使用 --help 查看用法）");
throw std::runtime_error(
FormatError("cli", "缺少输入文件：请提供 <input.sy>（使用 --help 查看用法）"));
}
if (!opt.emit_parse_tree && !opt.emit_ir && !opt.emit_asm) {
throw std::runtime_error(
"未选择任何输出：请使用 --emit-parse-tree / --emit-ir / --emit-asm");
throw std::runtime_error(FormatError(
"cli", "未选择任何输出：请使用 --emit-parse-tree / --emit-ir / --emit-asm"));
}
return opt;
}

42
src/utils/Log.cpp
View File

@@ -5,6 +5,48 @@
#include "utils/Log.h"
#include <ostream>
#include <string>
namespace {
bool IsCLIError(const std::string_view msg) {
return HasErrorPrefix(msg, "cli");
}
} // namespace
void LogInfo(const std::string_view msg, std::ostream& os) {
os << "[info] " << msg << "\n";
}
void LogError(const std::string_view msg, std::ostream& os) {
os << "[error] " << msg << "\n";
}
std::string FormatError(const std::string_view stage,
const std::string_view msg) {
return "[" + std::string(stage) + "] " + std::string(msg);
}
std::string FormatErrorAt(const std::string_view stage, const std::size_t line,
const std::size_t column,
const std::string_view msg) {
return "[" + std::string(stage) + "] @" + std::to_string(line) + ":" +
std::to_string(column) + " - " + std::string(msg);
}
bool HasErrorPrefix(const std::string_view msg, const std::string_view stage) {
const std::string prefix = "[" + std::string(stage) + "]";
return msg.rfind(prefix, 0) == 0;
}
void PrintException(std::ostream& os, const std::exception& ex) {
LogError(ex.what(), os);
if (IsCLIError(ex.what())) {
os << "\n";
PrintHelp(os);
}
}
void PrintHelp(std::ostream& os) {
os << "SysY Compiler (课程实验最小可运行示例)\n"

15
src/utils/Log.h
View File

@@ -1,11 +1,20 @@
// 轻量日志接口。
#pragma once
#include <cstddef>
#include <exception>
#include <iosfwd>
#include <iostream>
#include <string>
#include <string_view>
#define LOG_INFO(msg) std::cerr << "[info] " << msg << "\n"
#define LOG_ERROR(msg) std::cerr << "[error] " << msg << "\n"
void LogInfo(std::string_view msg, std::ostream& os);
void LogError(std::string_view msg, std::ostream& os);
std::string FormatError(std::string_view stage, std::string_view msg);
std::string FormatErrorAt(std::string_view stage, std::size_t line,
std::size_t column, std::string_view msg);
bool HasErrorPrefix(std::string_view msg, std::string_view stage);
void PrintException(std::ostream& os, const std::exception& ex);
// 打印命令行帮助信息（用于 `compiler --help`）。
void PrintHelp(std::ostream& os);