C++20 以 Bazel & Clang 开始

C++20 如何以 Bazel & Clang 进行构建呢？

本文将介绍：

Bazel 构建系统的安装
LLVM 编译系统的安装
- Clang is an "LLVM native" C/C++/Objective-C compiler
Bazel Clang 工具链的配置
C++20 库与应用的构建

本文示例可见： https://github.com/ikuokuo/st...

本文是于 Ubuntu 20 上进行的实践，Windows 可以用 WSL 准备环境。

安装 Bazel，以二进制方式

Bazelisk 是安装 Bazel 的推荐方式，我们安装它的二进制发布即可：

cd ~
wget https://github.com/bazelbuild/bazelisk/releases/download/v1.12.0/bazelisk-linux-amd64 -O bazelisk-1.12.0-linux-amd64
chmod a+x bazelisk-*

sudo ln -s $(pwd)/bazelisk-1.12.0-linux-amd64 /usr/local/bin/bazel

touch WORKSPACE
# 国内下载 Bazel 可能遇到如下问题，配置 .bazeliskrc 解决
# could not resolve the version 'latest' to an actual version number
#  https://github.com/bazelbuild/bazelisk/issues/220
cat <<-EOF > .bazeliskrc
BAZELISK_BASE_URL=https://github.com/bazelbuild/bazel/releases/download
USE_BAZEL_VERSION=5.2.0
EOF

bazel version

更多方式，可见官方文档。进一步，推荐安装 buildtools，下载后软链一下：

sudo ln -s $(pwd)/buildifier-5.1.0-linux-amd64 /usr/local/bin/buildifier
sudo ln -s $(pwd)/buildozer-5.1.0-linux-amd64 /usr/local/bin/buildozer

Bazel 如何构建 C++ 项目，可见我的 Start Bazel 笔记。

安装 LLVM，以源码方式

Clang 有关 std::fromat 文本格式化的特性，默认未开启：

The paper is implemented but still marked as an incomplete feature (the feature-test macro is not set and the libary is only available when built with LIBCXX_ENABLE_INCOMPLETE_FEATURES). Not yet implemented LWG-issues will cause API and ABI breakage.

C++20 特性，编译器支持情况：

因此，这里以源码方式安装 LLVM，需要构建 Clang & libc++：

git clone -b llvmorg-14.0.6 --depth 1 https://github.com/llvm/llvm-project.git

cd llvm-project
mkdir _build
cd _build

# llvm install path, such as /usr/local/llvm
LLVM_PREFIX=$HOME/Apps/llvm-14.0.6

cmake -DCMAKE_BUILD_TYPE=Release \
-DCMAKE_INSTALL_PREFIX=$LLVM_PREFIX \
-DLLVM_ENABLE_PROJECTS=clang \
-DLLVM_ENABLE_RUNTIMES="libcxx;libcxxabi" \
-DLIBCXX_ENABLE_INCOMPLETE_FEATURES=ON \
../llvm

make -j`nproc`
make install

sudo ln -s $LLVM_PREFIX /usr/local/llvm

cat <<-EOF >> ~/.bashrc
# llvm
export LLVM_HOME=/usr/local/llvm
export PATH=\$LLVM_HOME/bin:\$PATH
export LD_LIBRARY_PATH=\$LLVM_HOME/lib/x86_64-unknown-linux-gnu:\$LD_LIBRARY_PATH
EOF

llvm-config --version
clang --version

LLVM_PREFIX 安装路径自己决定。最后，编译测试：

cat <<-EOF > hello.cc
#include 
#include 

int main() {
  std::string message = std::format("The answer is {}.", 42);
  std::cout << message << std::endl;
}
EOF

clang++ -std=c++20 -stdlib=libc++ hello.cc -o hello

./hello

安装 LLVM，以二进制方式

可省略该节。本文实践未用此方式，因为想开启更多 C++20 特性。这里仅作记录，有需要可参考。

方式 1. 安装二进制发布：

cd ~
wget https://github.com/llvm/llvm-project/releases/download/llvmorg-13.0.0/clang+llvm-13.0.0-x86_64-linux-gnu-ubuntu-20.04.tar.xz
tar -xf clang+llvm-*.tar.xz

sudo ln -s $(pwd)/clang+llvm-13.0.0-x86_64-linux-gnu-ubuntu-20.04 /usr/local/llvm

cat <<-EOF >> ~/.bashrc
# llvm
export LLVM_HOME=/usr/local/llvm
export PATH=\$LLVM_HOME/bin:\$PATH
EOF

llvm-config --version
clang --version

方式 2. 用 apt 进行安装： https://apt.llvm.org/

方式 3. 用已配好的工具链： LLVM toolchain for Bazel

配置 Clang 工具链

本文依照 Bazel Tutorial: Configure C++ Toolchains 步骤配置的 Clang 工具链，最后项目根目录会多如下文件：

WORKSPACE 表示 Bazel 工作区，内容空。

.bazelrc 允许 --config=clang_config 启用 Clang 工具链：

# Use our custom-configured c++ toolchain.
build:clang_config --crosstool_top=//toolchain:clang_suite

# Use --cpu as a differentiator.
build:clang_config --cpu=linux_x86_64

# Use the default Bazel C++ toolchain to build the tools used during the build.
build:clang_config --host_crosstool_top=@bazel_tools//tools/cpp:toolchain

toolchain/BUILD 配置 Clang 工具链信息：

load(":cc_toolchain_config.bzl", "cc_toolchain_config")

package(default_visibility = ["//visibility:public"])

#filegroup(name = "clang_suite")

cc_toolchain_suite(
    name = "clang_suite",
    toolchains = {
        "linux_x86_64": ":linux_x86_64_toolchain",
    },
)

filegroup(name = "empty")

cc_toolchain(
    name = "linux_x86_64_toolchain",
    toolchain_identifier = "linux_x86_64-toolchain",
    toolchain_config = ":linux_x86_64_toolchain_config",
    all_files = ":empty",
    compiler_files = ":empty",
    dwp_files = ":empty",
    linker_files = ":empty",
    objcopy_files = ":empty",
    strip_files = ":empty",
    supports_param_files = 0,
)

#filegroup(name = "linux_x86_64_toolchain_config")

cc_toolchain_config(name = "linux_x86_64_toolchain_config")

toolchain/cc_toolchain_config.bzl 配置 Clang 工具链规则：

# C++ Toolchain Configuration
#  https://bazel.build/docs/cc-toolchain-config-reference
#  https://github.com/bazelbuild/bazel/blob/master/tools/build_defs/cc/action_names.bzl
load("@bazel_tools//tools/build_defs/cc:action_names.bzl", "ACTION_NAMES")
load(
    "@bazel_tools//tools/cpp:cc_toolchain_config_lib.bzl",
    "feature",
    "flag_group",
    "flag_set",
    "tool_path",
)

all_compile_actions = [
    ACTION_NAMES.c_compile,
    ACTION_NAMES.cpp_compile,
    ACTION_NAMES.linkstamp_compile,
    ACTION_NAMES.assemble,
    ACTION_NAMES.preprocess_assemble,
    ACTION_NAMES.cpp_header_parsing,
    ACTION_NAMES.cpp_module_compile,
    ACTION_NAMES.cpp_module_codegen,
]

all_link_actions = [
    ACTION_NAMES.cpp_link_executable,
    ACTION_NAMES.cpp_link_dynamic_library,
    ACTION_NAMES.cpp_link_nodeps_dynamic_library,
]

def _impl(ctx):
    llvm_version = "14.0.6"
    llvm_prefix = "/home/john/Apps/llvm-{}".format(llvm_version)
    llvm_bindir = llvm_prefix + "/bin"

    tool_paths = [
        tool_path(
            name = "gcc",
            path = llvm_bindir + "/clang",
        ),
        tool_path(
            name = "ld",
            path = llvm_bindir + "/ld.lld",
        ),
        tool_path(
            name = "ar",
            path = llvm_bindir + "/llvm-ar",
        ),
        tool_path(
            name = "cpp",
            path = llvm_bindir + "/clang-cpp",
        ),
        tool_path(
            name = "gcov",
            path = llvm_bindir + "/llvm-cov",
        ),
        tool_path(
            name = "nm",
            path = llvm_bindir + "/llvm-nm",
        ),
        tool_path(
            name = "objdump",
            path = llvm_bindir + "/llvm-objdump",
        ),
        tool_path(
            name = "strip",
            path = llvm_bindir + "/llvm-strip",
        ),
    ]

    features = [
        feature(
            name = "default_compiler_flags",
            enabled = True,
            flag_sets = [
                flag_set(
                    actions = all_compile_actions,
                    flag_groups = ([
                        flag_group(
                            flags = [
                                "-O2", "-DNDEBUG",
                                "-Wall", "-Wextra", "-Wpedantic", "-fPIC",
                                "-std=c++20", "-stdlib=libc++",
                            ],
                        ),
                    ]),
                ),
            ],
        ),
        feature(
            name = "default_linker_flags",
            enabled = True,
            flag_sets = [
                flag_set(
                    actions = all_link_actions,
                    flag_groups = ([
                        flag_group(
                            flags = [
                                "-lc++", "-lc++abi",
                                "-lm", "-ldl", "-lpthread",
                            ],
                        ),
                    ]),
                ),
            ],
        ),
    ]

    return cc_common.create_cc_toolchain_config_info(
        ctx = ctx,
        features = features,
        cxx_builtin_include_directories = [
            llvm_prefix + "/lib/clang/{}/include".format(llvm_version),
            llvm_prefix + "/include/x86_64-unknown-linux-gnu/c++/v1",
            llvm_prefix + "/include/c++/v1",
            "/usr/local/include",
            "/usr/include/x86_64-linux-gnu",
            "/usr/include",
        ],
        toolchain_identifier = "local",
        host_system_name = "local",
        target_system_name = "local",
        target_cpu = "linux_x86_64",
        target_libc = "unknown",
        compiler = "clang",
        abi_version = "unknown",
        abi_libc_version = "unknown",
        tool_paths = tool_paths,
    )

cc_toolchain_config = rule(
    implementation = _impl,
    attrs = {},
    provides = [CcToolchainConfigInfo],
)

llvm_prefix 给到自己的 LLVM 安装路径。

构建 C++20 库与应用

本文示例的 code/00/ 路径下准备了 C++20 的库与应用：

code/00/
├── BUILD
├── greet
│   ├── BUILD
│   ├── greet.cc
│   └── greet.h
└── main.cc

编写 binary

main.cc:

#include 
#include 
#include 
#include 

#include "greet/greet.h"

template 
std::string dyna_print(std::string_view rt_fmt_str, Args&&... args) {
  return std::vformat(rt_fmt_str, std::make_format_args(args...));
}

int main() {
  std::cout << greet::hello("world") << std::endl;

  std::string fmt;
  for (int i{}; i != 3; ++i) {
    fmt += "{} ";  // constructs the formatting string
    std::cout << fmt << " : ";
    std::cout << dyna_print(fmt, "alpha", 'Z', 3.14, "unused");
    std::cout << '\n';
  }
}

BUILD:

load("@rules_cc//cc:defs.bzl", "cc_binary")

cc_binary(
    name = "main",
    srcs = ["main.cc"],
    deps = [
        "//code/00/greet:greet",
    ],
)

编写 library

greet.h:

#pragma once

#include 
#include 

namespace greet {

std::string hello(std::string_view who);

}  // namespace greet

greet.cc:

#include "greet.h"

#include 
#include 

namespace greet {

std::string hello(std::string_view who) {
  return std::format("Hello {}!", std::move(who));
}

}  // namespace greet