mirror of
https://github.com/qicosmos/rest_rpc.git
synced 2026-08-29 16:40:48 +08:00
update readme
This commit is contained in:
qicosmos
@@ -7,7 +7,7 @@
|
||||
| macOS Monterey 12 (AppleClang 14.0.0.14000029) |  |
|
||||
| Windows Server 2022 (MSVC 19.33.31630.0) |  |
|
||||
|
||||
c++11, high performance, cross platform, easy to use rpc framework.
|
||||
c++20, high performance, cross platform, easy to use rpc framework.
|
||||
|
||||
It's so easy to love RPC.
|
||||
|
||||
@@ -15,7 +15,7 @@ Modern C++开发的RPC库就是这么简单好用！
|
||||
|
||||
# rest_rpc简介
|
||||
|
||||
rest_rpc是一个高性能、易用、跨平台、header only的c++11 rpc库，它的目标是让tcp通信变得非常简单易用，即使不懂网络通信的人也可以直接使用它。它依赖header-only的standalone [asio](https://github.com/chriskohlhoff/asio)(tag:asio-1-36-0)
|
||||
rest_rpc是一个高性能、易用、跨平台、header only的基于c++20 协程的rpc 库，它的目标是让tcp通信变得非常简单易用，即使不懂网络通信的人也可以直接使用它。它依赖header-only的standalone [asio](https://github.com/chriskohlhoff/asio)(tag:asio-1-36-0)
|
||||
|
||||
可以快速上手，使用者只需要关注自己的业务逻辑即可。
|
||||
|
||||
@@ -36,29 +36,36 @@ rest_rpc为用户提供了非常简单易用的接口，几行代码就可以实
|
||||
```
|
||||
//服务端注册加法rpc服务
|
||||
|
||||
struct dummy{
|
||||
int add(rpc_conn conn, int a, int b) { return a + b; }
|
||||
};
|
||||
int add(rpc_conn conn, int a, int b) { return a + b; }
|
||||
|
||||
int main(){
|
||||
rpc_server server(9000, std::thread::hardware_concurrency());
|
||||
rpc_server server("127.0.0.1:9004", std::thread::hardware_concurrency());
|
||||
|
||||
dummy d;
|
||||
server.register_handler("add", &dummy::add, &d);
|
||||
|
||||
server.run();
|
||||
server.register_handler<add>();
|
||||
|
||||
server.start();
|
||||
}
|
||||
```
|
||||
|
||||
```
|
||||
//客户端调用加法的rpc服务
|
||||
int main(){
|
||||
rpc_client client("127.0.0.1", 9000);
|
||||
client.connect();
|
||||
|
||||
int result = client.call<int>("add", 1, 2);
|
||||
|
||||
client.run();
|
||||
auto rpc_call = []() -> asio::awaitable<void> {
|
||||
rpc_client client;
|
||||
auto ec = co_await client.connect("127.0.0.1:9004");
|
||||
if(ec) {
|
||||
REST_LOG_ERROR << ec0.message();
|
||||
co_return;
|
||||
}
|
||||
|
||||
auto r = co_await client.call<add>(1, 2);
|
||||
if(r.ec==rpc_errc::ok) {
|
||||
REST_LOG_INFO << "call result: " << r.value;
|
||||
assert(r.value == 3);
|
||||
}
|
||||
};
|
||||
|
||||
sync_wait(get_global_executor(), rpc_call());
|
||||
}
|
||||
```
|
||||
|
||||
@@ -69,133 +76,134 @@ int main(){
|
||||
|
||||
//1.先定义person对象
|
||||
struct person {
|
||||
int id;
|
||||
std::string name;
|
||||
int age;
|
||||
|
||||
MSGPACK_DEFINE(id, name, age);
|
||||
int id;
|
||||
std::string name;
|
||||
int age;
|
||||
};
|
||||
|
||||
//2.提供并服务
|
||||
person get_person(rpc_conn conn) {
|
||||
return { 1, "tom", 20 };
|
||||
person get_person(person p) {
|
||||
p.name = "jack";
|
||||
return p;
|
||||
}
|
||||
|
||||
int main(){
|
||||
//...
|
||||
server.register_handler("get_person", get_person);
|
||||
rpc_server server("127.0.0.1:9004", std::thread::hardware_concurrency());
|
||||
server.register_handler<get_person>();
|
||||
server.start();
|
||||
}
|
||||
```
|
||||
|
||||
```
|
||||
//客户端调用获取person对象的rpc服务
|
||||
int main(){
|
||||
rpc_client client("127.0.0.1", 9000);
|
||||
client.connect();
|
||||
|
||||
person result = client.call<person>("get_person");
|
||||
std::cout << result.name << std::endl;
|
||||
|
||||
client.run();
|
||||
}
|
||||
```
|
||||
|
||||
## 酷
|
||||
|
||||
异步？
|
||||
|
||||
同步？
|
||||
|
||||
future?
|
||||
|
||||
callback？
|
||||
|
||||
当初为了提供什么样的接口在社区群里还争论了一番，有人希望提供callback接口，有人希望提供future接口，最后我
|
||||
决定都提供，专治强迫症患者:)
|
||||
|
||||
现在想要的这些接口都给你提供了，你想用什么类型的接口就用什么类型的接口，够酷吧，让我们来看看怎么用这些接口吧：
|
||||
|
||||
```
|
||||
//服务端提供echo服务
|
||||
std::string echo(rpc_conn conn, const std::string& src) {
|
||||
return src;
|
||||
}
|
||||
|
||||
server.register_handler("echo", echo);
|
||||
```
|
||||
|
||||
客户端同步接口
|
||||
|
||||
```
|
||||
auto result = client.call<std::string>("echo", "hello");
|
||||
```
|
||||
|
||||
客户端异步回调接口
|
||||
|
||||
```
|
||||
client.async_call<R>("echo", [](asio::error_code ec, R data){
|
||||
std::cout << "echo " << data << '\n';
|
||||
});
|
||||
```
|
||||
|
||||
## async_call接口说明
|
||||
有两个重载的async_call接口，一个是返回future的接口，一个是带超时的异步接口。
|
||||
|
||||
返回future的async_call接口：
|
||||
```
|
||||
std::future<std::string> future = client.async_call<CallModel::future>("echo", "purecpp");
|
||||
```
|
||||
|
||||
带超时的异步回调接口：
|
||||
```
|
||||
async_call<R, timeout_ms>("some_rpc_service_name", callback, service_args...);
|
||||
```
|
||||
|
||||
如果不显式设置超时时间的话，则会用默认的5s超时.
|
||||
```
|
||||
async_call<R>("some_rpc_service_name", callback, args...);
|
||||
```
|
||||
|
||||
```
|
||||
client.async_call<std::string>("echo", [](asio::error_code ec, std::string result) {
|
||||
if (ec) {
|
||||
std::cout << ec.message() <<" "<< data << "\n";
|
||||
return;
|
||||
auto rpc_call = []() -> asio::awaitable<void> {
|
||||
rpc_client client;
|
||||
auto ec = co_await client.connect("127.0.0.1:9004");
|
||||
if(ec) {
|
||||
REST_LOG_ERROR << ec0.message();
|
||||
co_return;
|
||||
}
|
||||
|
||||
std::cout << result << " async\n";
|
||||
}, "purecpp");
|
||||
```
|
||||
|
||||
客户端异步future接口
|
||||
|
||||
```
|
||||
auto future = client->async_call<FUTURE>("echo", "hello");
|
||||
auto status = future.wait_for(std::chrono::seconds(2));
|
||||
if (status == std::future_status::timeout) {
|
||||
std::cout << "timeout\n";
|
||||
}
|
||||
else if (status == std::future_status::ready) {
|
||||
auto str = future.get().as<std::string>();
|
||||
std::cout << "echo " << str << '\n';
|
||||
|
||||
person p{1, "tom", 20};
|
||||
auto r = co_await client.call<get_person>(p);
|
||||
if(r.ec==rpc_errc::ok) {
|
||||
REST_LOG_INFO << "call result: " << r.value.name;
|
||||
assert(r.value.name == "jack");
|
||||
}
|
||||
};
|
||||
|
||||
sync_wait(get_global_executor(), rpc_call());
|
||||
}
|
||||
```
|
||||
|
||||
除了上面的这些很棒的接口之外，更酷的是rest_rpc还支持了订阅发布的功能，这是目前很多rpc库做不到的。
|
||||
## 发布订阅
|
||||
以订阅某个topic为例：
|
||||
|
||||
服务端订阅发布的例子在这里：
|
||||
server 端代码：
|
||||
```cpp
|
||||
void publish() {
|
||||
rpc_server server("127.0.0.1:9004", 4);
|
||||
server.async_start();
|
||||
|
||||
REST_LOG_INFO << "will pubish, waiting for input";
|
||||
auto pub = [&]() -> asio::awaitable<void> {
|
||||
std::string str;
|
||||
while (true) {
|
||||
std::cin >> str;
|
||||
if(str == "quit") {
|
||||
break;
|
||||
}
|
||||
|
||||
co_await server.publish("topic1", str);// 向客户端发布一个string，你也可以发布一个对象，内部会自动序列化
|
||||
}
|
||||
};
|
||||
|
||||
sync_wait(get_global_executor(), pub());
|
||||
}
|
||||
|
||||
https://github.com/qicosmos/rest_rpc/blob/master/examples/server/main.cpp#L121
|
||||
https://github.com/qicosmos/rest_rpc/blob/master/examples/client/main.cpp#L383
|
||||
client 端代码：
|
||||
```cpp
|
||||
void subscribe() {
|
||||
REST_LOG_INFO << "will subscribe, waiting for publish";
|
||||
auto sub = [&]() -> asio::awaitable<void> {
|
||||
rpc_client client;
|
||||
co_await client.connect("127.0.0.1:9004");
|
||||
while (true) {
|
||||
// 订阅topic1，会自动将结果反序列化为std::string, 如果publish是一个person对象，则subscribe参数填person，内部会自动反序列化
|
||||
auto [ec, result] = co_await client.subscribe<std::string>("topic1");
|
||||
if (ec != rpc_errc::ok) {
|
||||
REST_LOG_ERROR << "subscribe failed: " << make_error_code(ec).message();
|
||||
break;
|
||||
}
|
||||
|
||||
REST_LOG_INFO << result;
|
||||
}
|
||||
};
|
||||
|
||||
sync_wait(get_global_executor(), sub());
|
||||
}
|
||||
```
|
||||
|
||||
## 快
|
||||
|
||||
rest_rpc是目前最快的rpc库，具体和grpc和brpc做了性能对比测试，rest_rpc性能是最高的，远超grpc。
|
||||
|
||||
性能测试的结果在这里：
|
||||
性能测试代码在这里：
|
||||
|
||||
https://github.com/qicosmos/rest_rpc/blob/master/doc/%E5%8D%95%E6%9C%BA%E4%B8%8Arest_rpc%E5%92%8Cbrpc%E6%80%A7%E8%83%BD%E6%B5%8B%E8%AF%95.md
|
||||
https://github.com/qicosmos/rest_rpc/tree/master/tests/bench.cpp
|
||||
|
||||
## 使用自己的序列化库
|
||||
rest_rpc 默认使用yalantinglibs的struct_pack 去做系列化/反序列化的，它的性能非常好。
|
||||
|
||||
rest_rpc 也支持用户使用自己的序列化库，只需要去实现一个序列化和一个反序列化函数。
|
||||
```cpp
|
||||
namespace user_codec {
|
||||
// adl lookup in user_codec namespace
|
||||
template <typename... Args>
|
||||
std::string serialize(rest_adl_tag, Args &&...args) {
|
||||
msgpack::sbuffer buffer(2 * 1024);
|
||||
if constexpr (sizeof...(Args) > 1) {
|
||||
msgpack::pack(buffer, std::forward_as_tuple(std::forward<Args>(args)...));
|
||||
} else {
|
||||
msgpack::pack(buffer, std::forward<Args>(args)...);
|
||||
}
|
||||
|
||||
return std::string(buffer.data(), buffer.size());
|
||||
}
|
||||
|
||||
template <typename T> T deserialize(rest_adl_tag, std::string_view data) {
|
||||
try {
|
||||
static msgpack::unpacked msg;
|
||||
msgpack::unpack(msg, data.data(), data.size());
|
||||
return msg.get().as<T>();
|
||||
} catch (...) {
|
||||
return T{};
|
||||
}
|
||||
}
|
||||
} // namespace user_codec
|
||||
```
|
||||
实现这两个函数之后rest_rpc 将会使用自定义的序列化/反序列化函数了。
|
||||
|
||||
# rest_rpc的更多用法
|
||||
|
||||
@@ -203,9 +211,6 @@ https://github.com/qicosmos/rest_rpc/blob/master/doc/%E5%8D%95%E6%9C%BA%E4%B8%8A
|
||||
|
||||
https://github.com/qicosmos/rest_rpc/tree/master/examples
|
||||
|
||||
# future
|
||||
|
||||
make an IDL tool to genrate the client code.
|
||||
|
||||
## 社区和群
|
||||
purecpp.cn
|
||||
|
||||
Reference in New Issue
Block a user