Boost.Redis

Boost.Redis is a high-level Redis client library built on top of Boost.Asio that implements the Redis protocol RESP3.

Full documentation is here.

Requirements

The requirements for using Boost.Redis are:

• Boost 1.84 or higher. Boost.Redis is included in Boost installations since Boost 1.84.
• C++17 or higher. Supported compilers include gcc 11 and later, clang 11 and later, and Visual Studio 16 (2019) and later.
• Redis 6 or higher (must support RESP3).
• OpenSSL.

The documentation assumes basic-level knowledge about Redis and Boost.Asio.

Building the library

To use the library it is necessary to include the following:

#include <boost/redis/src.hpp>

in exactly one source file in your applications. Otherwise, the library is header-only.

Boost.Redis unconditionally requires OpenSSL. Targets using Boost.Redis need to link to the OpenSSL libraries.

Tutorial

The code below uses a short-lived connection to ping a Redis server:

#include <boost/redis/connection.hpp>
#include <boost/asio/co_spawn.hpp>
#include <boost/asio/consign.hpp>
#include <boost/asio/detached.hpp>
#include <iostream>

namespace net = boost::asio;
using boost::redis::request;
using boost::redis::response;
using boost::redis::config;
using boost::redis::connection;

auto co_main(config const& cfg) -> net::awaitable<void>
{
   auto conn = std::make_shared<connection>(co_await net::this_coro::executor);
   conn->async_run(cfg, {}, net::consign(net::detached, conn));

   // A request containing only a ping command.
   request req;
   req.push("PING", "Hello world");

   // Response object.
   response<std::string> resp;

   // Executes the request.
   co_await conn->async_exec(req, resp);
   conn->cancel();

   std::cout << "PING: " << std::get<0>(resp).value() << std::endl;
}

The roles played by the async_run and async_exec functions are:

• connection::async_exec: executes the commands contained in the request and stores the individual responses in the response object. Can be called from multiple places in your code concurrently.
• connection::async_run: keeps the connection healthy. It takes care of hostname resolution, session establishment, health-checks, reconnection and coordination of low-level read and write operations. It should be called only once per connection, regardless of the number of requests to execute.

Server pushes

Redis servers can also send a variety of pushes to the client. Some of them are:

• Pubsub messages.
• Keyspace notifications.
• Client-side caching.

The connection class supports server pushes by means of the connection::async_receive function, which can be called in the same connection that is being used to execute commands. The coroutine below shows how to use it:

auto
receiver(std::shared_ptr<connection> conn) -> net::awaitable<void>
{
   request req;
   req.push("SUBSCRIBE", "channel");

   generic_response resp;
   conn->set_receive_response(resp);

   // Loop while reconnection is enabled
   while (conn->will_reconnect()) {

      // Reconnect to channels.
      co_await conn->async_exec(req, ignore);

      // Loop reading Redis pushes.
      for (;;) {
         error_code ec;
         co_await conn->async_receive(resp, net::redirect_error(net::use_awaitable, ec));
         if (ec)
            break; // Connection lost, break so we can reconnect to channels.

         // Use the response resp in some way and then clear it.
         ...

         consume_one(resp);
      }
   }
}

Further reading

Full documentation is here.