c-compiler/src/main.cpp

#include <llvm/IR/LLVMContext.h>
#include <llvm/ADT/StringRef.h>
#include <llvm/Support/TargetSelect.h>
#include "llvm/Target/TargetMachine.h"
#include "llvm/Target/TargetOptions.h"
#include <llvm/MC/TargetRegistry.h>
#include <llvm/IR/IRBuilder.h>
#include <llvm/IR/Module.h>
#include <llvm/TargetParser/Host.h>
#include <llvm/Support/raw_ostream.h>
#include <llvm/Support/FileSystem.h>
#include <llvm/IR/LegacyPassManager.h>
#include <llvm/IR/Verifier.h>

#include <iostream>
#include <fstream>
#include <cctype>
#include <string>

#include "tokens.h"
#include "parsing.h"
#include "errors.h"

std::string read_file(std::string_view filepath) {
    std::ifstream input{ std::string{filepath} };
    if (!input) {
        std::cerr << "Failed to read " << filepath << std::endl;
        return {};
    }

    std::string out{};
    std::string currLine;
    while (std::getline(input, currLine)) {
        if (out.length() > 0) {
            out = out + "\n";
        }
        out = out + currLine;
    }

    return out;
}

int write_file(std::string_view filepath, std::string& content) {
    std::ofstream output{ std::string{filepath} };
    if (!output) {
        std::cerr << "Failed to open " << filepath << " for reading" << std::endl;
        return 1;
    }

    output << content << std::endl;
    return 0;
}

struct CompileOutput {
    std::string llvm_ir_string;
    std::string obj_string;
};

/// @brief Compiles the contents from the given filename and returns output as strings
std::optional<CompileOutput> compile(std::string_view in_filename) {

    // Read contents of the file, print contents
    std::string out{ read_file(in_filename) };
    std::cout << out << std::endl;

    // Tokenize contents, print tokens
    auto tokens = token::tokenize(out, std::string{ in_filename });
    for (token::Token token : tokens) {
        std::cout << token << std::endl;
    }

    // Parse tokens
    auto stream = token::TokenStream{ tokens };
    std::vector<std::unique_ptr<AST::TopLevelStatement>> statements;
    auto statement = parsing::parse_top_level_statement(stream);
    while (statement.ok()) {
        statements.push_back(statement.unwrap());
        statement = parsing::parse_top_level_statement(stream);
    }
    if (stream.peek().type != token::Type::Eof) {
        std::cerr << statement.unwrap_err() << std::endl;
        return {};
    }
    stream.expect(token::Type::Eof);

    // Prepare compiler
    auto LLVMContext = std::make_unique<llvm::LLVMContext>();
    auto LLVMModule = std::make_unique<llvm::Module>("test.c", *LLVMContext);
    auto LLVMBuilder = std::make_unique<llvm::IRBuilder<>>(*LLVMContext);

    codegen::Builder builder{
        std::move(LLVMContext),
        std::move(LLVMModule),
        std::move(LLVMBuilder),
        {}
    };

    // Perform static analysis

    typecheck::State typecheck_state{};
    typecheck_state.binops = types::create_binops();
    typecheck_state.casts = types::create_casts();

    typecheck::Scope typecheck_scope{};
    for (auto& tls : statements) {
        std::cout << tls->formatted() << std::endl;
        tls->typecheck(typecheck_state, typecheck_scope);
    }

    if (typecheck_state.errors.size() > 0) {
        std::cerr << "Errors while typechecking:" << std::endl;
        for (auto& error : typecheck_state.errors) {
            std::cerr << " " << error.what() << std::endl;
            std::cerr << "   at " << error.m_meta.start.line + 1 << ":" << error.m_meta.start.col + 1;
            std::cerr << " to " << error.m_meta.end.line + 1 << ":" << error.m_meta.end.col + 1 << std::endl;
        }
        return {};
    }

    codegen::Scope cg_scope{
        .binops = typecheck_state.binops,
        .casts = typecheck_state.casts,
        .values = {},
        .is_lvalue = false,
    };

    // Compile parsed output
    try {
        for (auto& tls : statements) {
            std::cout << tls->formatted() << std::endl;
            tls->codegen(builder, cg_scope);
        }
    }
    catch (CompileError& error) {
        std::cerr << "FATAL: " << error.what() << std::endl;
        std::cerr << "   at " << error.m_meta.start.line + 1 << ":" << error.m_meta.start.col + 1;
        std::cerr << " to " << error.m_meta.end.line + 1 << ":" << error.m_meta.end.col + 1 << std::endl;
        return {};
    }

    // Prepare to print module as output
    llvm::InitializeAllTargetInfos();
    llvm::InitializeAllTargets();
    llvm::InitializeAllTargetMCs();
    llvm::InitializeAllAsmParsers();
    llvm::InitializeAllAsmPrinters();

    auto TargetTriple = llvm::sys::getDefaultTargetTriple();
    std::string Error;
    auto Target = llvm::TargetRegistry::lookupTarget(TargetTriple, Error);

    llvm::TargetOptions opt;
    auto TargetMachine = Target->createTargetMachine(llvm::Triple(TargetTriple), "generic", "", opt, llvm::Reloc::PIC_);
    builder.mod->setDataLayout(TargetMachine->createDataLayout());
    builder.mod->setTargetTriple(TargetMachine->getTargetTriple());

    // Print output to string
    std::string llvm_ir_string;
    llvm::raw_string_ostream llvm_ir_dest{ llvm_ir_string };

    builder.mod->print(llvm_ir_dest, nullptr);
    llvm_ir_dest.flush();

    // Print output to obj-file
    std::error_code EC;
    std::string obj_string;
    llvm::raw_string_ostream obj_stream{ obj_string };
    llvm::buffer_ostream obj_dest{ obj_stream };

    if (EC) {
        llvm::errs() << "Unable to open file (" << EC.message() << ")";
        return {};
    }

    llvm::legacy::PassManager pass{};
    if (TargetMachine->addPassesToEmitFile(pass, obj_dest, nullptr, llvm::CodeGenFileType::ObjectFile)) {
        llvm::errs() << "Target Machine can't emit file";
        return {};
    }

    pass.run(*builder.mod);
    obj_string = (std::string_view)obj_dest.str();

    return CompileOutput{
        llvm_ir_string,
        obj_string,
    };
}

struct ClosePipeDeleter {
    // Note 2: Consider adding noexcept.
    void operator()(FILE* file) const {
        pclose(file);
    }
};

/// @brief Executes a given command and returns the output as std::string
std::string exec(const char* cmd) {
    std::array<char, 128> buffer;
    std::string result;
    std::unique_ptr<FILE, ClosePipeDeleter> pipe(popen(cmd, "r"), ClosePipeDeleter{});
    if (!pipe) {
        throw std::runtime_error("popen() failed!");
    }
    while (fgets(buffer.data(), static_cast<int>(buffer.size()), pipe.get()) != nullptr) {
        result += buffer.data();
    }
    return result;
}

/// @brief Discovers the path of a given library using POSIX-util "whereis"
/// @return path to the library, empty string if nothing was found
std::string find_lib(std::string libname) {
    std::string cmd = "whereis " + libname;
    auto output = exec(cmd.c_str());
    output.erase(0, libname.size() + 2);
    auto path = output.substr(0, output.find('\n'));
    path = path.substr(0, path.find(' '));

    return path;
}

int main() {

    std::string in_filename{ "test.c" };
    std::string out_filename{ "test.o" };
    std::string exec_filename{ "test" };

    auto out = compile(in_filename);
    if (out) {
        // Print LLVM IR, produce Obj-file
        std::cout << out->llvm_ir_string << std::endl;
        write_file(out_filename, out->obj_string);

        // Find necessary libraries

        std::string command = "ld";
        auto linker = find_lib("ld-linux-x86-64.so.2");
        auto crt1 = find_lib("crt1.o");
        auto crti = find_lib("crti.o");
        auto crtn = find_lib("crtn.o");


        // Link everything together and produce a file with exec_filename
        std::string cmd = command + " -dynamic-linker " + linker + " -lc "
            + crt1 + " " + crti + " " + crtn + " "
            + out_filename + " -o " + exec_filename;
        std::cout << cmd << std::endl;
        std::cout << exec(cmd.c_str()) << std::endl;
    }

    return 0;
}