1591892489
JavaScript tutorial 42 - Includes method
This tutorial covers includes method functionality. It also provides a multiple examples using this method.
#javascript
What is GEEK
Buddha Community
1651147200
Ultra Fast and Low Latency Asynchronous Socket Server & Client C++
CppServer
Ultra fast and low latency asynchronous socket server & client C++ library with support TCP, SSL, UDP, HTTP, HTTPS, WebSocket protocols and 10K connections problem solution.
Has integration with high-level message protocol based on Fast Binary Encoding
Contents
- Features
- Requirements
- How to build?
- Examples
- Example: Asio service
- Example: Asio timer
- Example: TCP chat server
- Example: TCP chat client
- Example: SSL chat server
- Example: SSL chat client
- Example: UDP echo server
- Example: UDP echo client
- Example: UDP multicast server
- Example: UDP multicast client
- Example: Simple protocol
- Example: Simple protocol server
- Example: Simple protocol client
- Example: HTTP server
- Example: HTTP client
- Example: HTTPS server
- Example: HTTPS client
- Example: WebSocket chat server
- Example: WebSocket chat client
- Example: WebSocket secure chat server
- Example: WebSocket secure chat client
- Performance
- OpenSSL certificates
Features
- Cross platform (Linux, MacOS, Windows)
- Asynchronous communication
- Supported CPU scalability designs: IO service per thread, thread pool
- Supported transport protocols: TCP, SSL, UDP, UDP multicast
- Supported Web protocols: HTTP, HTTPS, WebSocket, WebSocket secure
- Supported Swagger OpenAPI iterative documentation
- Supported message protocol based on Fast Binary Encoding
Requirements
Optional:
How to build?
Linux: install required packages
sudo apt-get install -y binutils-dev uuid-dev libssl-dev
Install gil (git links) tool
pip3 install gil
Setup repository
git clone https://github.com/chronoxor/CppServer.git
cd CppServer
gil update
Linux
cd build
./unix.sh
MacOS
cd build
./unix.sh
Windows (MSYS2)
cd build
unix.bat
Windows (MinGW)
cd build
mingw.bat
Windows (Visual Studio)
cd build
vs.bat
Examples
Example: Asio service
Asio service is used to host all clients/servers based on Asio C++ library. It is implemented based on Asio C++ Library and use a separate thread to perform all asynchronous IO operations and communications.
The common usecase is to instantiate one Asio service, start the service and attach TCP/UDP/WebSocket servers or/and clients to it. One Asio service can handle several servers and clients asynchronously at the same time in one I/O thread. If you want to scale your servers or clients it is possible to create and use more than one Asio services to handle your servers/clients in balance.
Also it is possible to dispatch or post your custom handler into I/O thread. Dispatch will execute the handler immediately if the current thread is I/O one. Otherwise the handler will be enqueued to the I/O queue. In opposite the post method will always enqueue the handler into the I/O queue.
Here comes an example of using custom Asio service with dispatch/post methods:
#include "server/asio/service.h"
#include "threads/thread.h"
#include <iostream>
int main(int argc, char** argv)
{
// Create a new Asio service
auto service = std::make_shared<CppServer::Asio::Service>();
// Start the Asio service
std::cout << "Asio service starting...";
service->Start();
std::cout << "Done!" << std::endl;
// Dispatch
std::cout << "1 - Dispatch from the main thread with Id " << CppCommon::Thread::CurrentThreadId() << std::endl;
service->Dispatch([service]()
{
std::cout << "1.1 - Dispatched in thread with Id " << CppCommon::Thread::CurrentThreadId() << std::endl;
std::cout << "1.2 - Dispatch from thread with Id " << CppCommon::Thread::CurrentThreadId() << std::endl;
service->Dispatch([service]()
{
std::cout << "1.2.1 - Dispatched in thread with Id " << CppCommon::Thread::CurrentThreadId() << std::endl;
});
std::cout << "1.3 - Post from thread with Id " << CppCommon::Thread::CurrentThreadId() << std::endl;
service->Post([service]()
{
std::cout << "1.3.1 - Posted in thread with Id " << CppCommon::Thread::CurrentThreadId() << std::endl;
});
});
// Post
std::cout << "2 - Post from the main thread with Id " << CppCommon::Thread::CurrentThreadId() << std::endl;
service->Post([service]()
{
std::cout << "2.1 - Posted in thread with Id " << CppCommon::Thread::CurrentThreadId() << std::endl;
std::cout << "2.2 - Dispatch from thread with Id " << CppCommon::Thread::CurrentThreadId() << std::endl;
service->Dispatch([service]()
{
std::cout << "2.2.1 - Dispatched in thread with Id " << CppCommon::Thread::CurrentThreadId() << std::endl;
});
std::cout << "2.3 - Post from thread with Id " << CppCommon::Thread::CurrentThreadId() << std::endl;
service->Post([service]()
{
std::cout << "2.3.1 - Posted in thread with Id " << CppCommon::Thread::CurrentThreadId() << std::endl;
});
});
// Wait for a while...
CppCommon::Thread::Sleep(1000);
// Stop the Asio service
std::cout << "Asio service stopping...";
service->Stop();
std::cout << "Done!" << std::endl;
return 0;
}
Output of the above example is the following:
Asio service started!
1 - Dispatch from the main thread with Id 16744
2 - Post from the main thread with Id 16744
1.1 - Dispatched in thread with Id 19920
1.2 - Dispatch from thread with Id 19920
1.2.1 - Dispatched in thread with Id 19920
1.3 - Post from thread with Id 19920
2.1 - Posted in thread with Id 19920
2.2 - Dispatch from thread with Id 19920
2.2.1 - Dispatched in thread with Id 19920
2.3 - Post from thread with Id 19920
1.3.1 - Posted in thread with Id 19920
2.3.1 - Posted in thread with Id 19920
Asio service stopped!
Example: Asio timer
Here comes the example of Asio timer. It can be used to wait for some action in future with providing absolute time or relative time span. Asio timer can be used in synchronous or asynchronous modes.
#include "server/asio/timer.h"
#include "threads/thread.h"
#include <iostream>
class AsioTimer : public CppServer::Asio::Timer
{
public:
using CppServer::Asio::Timer::Timer;
protected:
void onTimer(bool canceled) override
{
std::cout << "Asio timer " << (canceled ? "canceled" : "expired") << std::endl;
}
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "Asio timer caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
};
int main(int argc, char** argv)
{
// Create a new Asio service
auto service = std::make_shared<CppServer::Asio::Service>();
// Start the Asio service
std::cout << "Asio service starting...";
service->Start();
std::cout << "Done!" << std::endl;
// Create a new Asio timer
auto timer = std::make_shared<AsioTimer>(service);
// Setup and synchronously wait for the timer
timer->Setup(CppCommon::UtcTime() + CppCommon::Timespan::seconds(1));
timer->WaitSync();
// Setup and asynchronously wait for the timer
timer->Setup(CppCommon::Timespan::seconds(1));
timer->WaitAsync();
// Wait for a while...
CppCommon::Thread::Sleep(2000);
// Setup and asynchronously wait for the timer
timer->Setup(CppCommon::Timespan::seconds(1));
timer->WaitAsync();
// Wait for a while...
CppCommon::Thread::Sleep(500);
// Cancel the timer
timer->Cancel();
// Wait for a while...
CppCommon::Thread::Sleep(500);
// Stop the Asio service
std::cout << "Asio service stopping...";
service->Stop();
std::cout << "Done!" << std::endl;
return 0;
}
Output of the above example is the following:
Asio service starting...Done!
Timer was expired
Timer was canceled
Asio service stopping...Done!
Example: TCP chat server
Here comes the example of the TCP chat server. It handles multiple TCP client sessions and multicast received message from any session to all ones. Also it is possible to send admin message directly from the server.
#include "server/asio/tcp_server.h"
#include "threads/thread.h"
#include <iostream>
class ChatSession : public CppServer::Asio::TCPSession
{
public:
using CppServer::Asio::TCPSession::TCPSession;
protected:
void onConnected() override
{
std::cout << "Chat TCP session with Id " << id() << " connected!" << std::endl;
// Send invite message
std::string message("Hello from TCP chat! Please send a message or '!' to disconnect the client!");
SendAsync(message);
}
void onDisconnected() override
{
std::cout << "Chat TCP session with Id " << id() << " disconnected!" << std::endl;
}
void onReceived(const void* buffer, size_t size) override
{
std::string message((const char*)buffer, size);
std::cout << "Incoming: " << message << std::endl;
// Multicast message to all connected sessions
server()->Multicast(message);
// If the buffer starts with '!' the disconnect the current session
if (message == "!")
DisconnectAsync();
}
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "Chat TCP session caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
};
class ChatServer : public CppServer::Asio::TCPServer
{
public:
using CppServer::Asio::TCPServer::TCPServer;
protected:
std::shared_ptr<CppServer::Asio::TCPSession> CreateSession(std::shared_ptr<CppServer::Asio::TCPServer> server) override
{
return std::make_shared<ChatSession>(server);
}
protected:
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "Chat TCP server caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
};
int main(int argc, char** argv)
{
// TCP server port
int port = 1111;
if (argc > 1)
port = std::atoi(argv[1]);
std::cout << "TCP server port: " << port << std::endl;
// Create a new Asio service
auto service = std::make_shared<CppServer::Asio::Service>();
// Start the Asio service
std::cout << "Asio service starting...";
service->Start();
std::cout << "Done!" << std::endl;
// Create a new TCP chat server
auto server = std::make_shared<ChatServer>(service, port);
// Start the server
std::cout << "Server starting...";
server->Start();
std::cout << "Done!" << std::endl;
std::cout << "Press Enter to stop the server or '!' to restart the server..." << std::endl;
// Perform text input
std::string line;
while (getline(std::cin, line))
{
if (line.empty())
break;
// Restart the server
if (line == "!")
{
std::cout << "Server restarting...";
server->Restart();
std::cout << "Done!" << std::endl;
continue;
}
// Multicast admin message to all sessions
line = "(admin) " + line;
server->Multicast(line);
}
// Stop the server
std::cout << "Server stopping...";
server->Stop();
std::cout << "Done!" << std::endl;
// Stop the Asio service
std::cout << "Asio service stopping...";
service->Stop();
std::cout << "Done!" << std::endl;
return 0;
}
Example: TCP chat client
Here comes the example of the TCP chat client. It connects to the TCP chat server and allows to send message to it and receive new messages.
#include "server/asio/tcp_client.h"
#include "threads/thread.h"
#include <atomic>
#include <iostream>
class ChatClient : public CppServer::Asio::TCPClient
{
public:
ChatClient(std::shared_ptr<CppServer::Asio::Service> service, const std::string& address, int port)
: CppServer::Asio::TCPClient(service, address, port)
{
_stop = false;
}
void DisconnectAndStop()
{
_stop = true;
DisconnectAsync();
while (IsConnected())
CppCommon::Thread::Yield();
}
protected:
void onConnected() override
{
std::cout << "Chat TCP client connected a new session with Id " << id() << std::endl;
}
void onDisconnected() override
{
std::cout << "Chat TCP client disconnected a session with Id " << id() << std::endl;
// Wait for a while...
CppCommon::Thread::Sleep(1000);
// Try to connect again
if (!_stop)
ConnectAsync();
}
void onReceived(const void* buffer, size_t size) override
{
std::cout << "Incoming: " << std::string((const char*)buffer, size) << std::endl;
}
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "Chat TCP client caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
private:
std::atomic<bool> _stop;
};
int main(int argc, char** argv)
{
// TCP server address
std::string address = "127.0.0.1";
if (argc > 1)
address = argv[1];
// TCP server port
int port = 1111;
if (argc > 2)
port = std::atoi(argv[2]);
std::cout << "TCP server address: " << address << std::endl;
std::cout << "TCP server port: " << port << std::endl;
// Create a new Asio service
auto service = std::make_shared<CppServer::Asio::Service>();
// Start the Asio service
std::cout << "Asio service starting...";
service->Start();
std::cout << "Done!" << std::endl;
// Create a new TCP chat client
auto client = std::make_shared<ChatClient>(service, address, port);
// Connect the client
std::cout << "Client connecting...";
client->ConnectAsync();
std::cout << "Done!" << std::endl;
std::cout << "Press Enter to stop the client or '!' to reconnect the client..." << std::endl;
// Perform text input
std::string line;
while (getline(std::cin, line))
{
if (line.empty())
break;
// Disconnect the client
if (line == "!")
{
std::cout << "Client disconnecting...";
client->DisconnectAsync();
std::cout << "Done!" << std::endl;
continue;
}
// Send the entered text to the chat server
client->SendAsync(line);
}
// Disconnect the client
std::cout << "Client disconnecting...";
client->DisconnectAndStop();
std::cout << "Done!" << std::endl;
// Stop the Asio service
std::cout << "Asio service stopping...";
service->Stop();
std::cout << "Done!" << std::endl;
return 0;
}
Example: SSL chat server
Here comes the example of the SSL chat server. It handles multiple SSL client sessions and multicast received message from any session to all ones. Also it is possible to send admin message directly from the server.
This example is very similar to the TCP one except the code that prepares SSL context and handshake handler.
#include "server/asio/ssl_server.h"
#include "threads/thread.h"
#include <iostream>
class ChatSession : public CppServer::Asio::SSLSession
{
public:
using CppServer::Asio::SSLSession::SSLSession;
protected:
void onConnected() override
{
std::cout << "Chat SSL session with Id " << id() << " connected!" << std::endl;
}
void onHandshaked() override
{
std::cout << "Chat SSL session with Id " << id() << " handshaked!" << std::endl;
// Send invite message
std::string message("Hello from SSL chat! Please send a message or '!' to disconnect the client!");
SendAsync(message.data(), message.size());
}
void onDisconnected() override
{
std::cout << "Chat SSL session with Id " << id() << " disconnected!" << std::endl;
}
void onReceived(const void* buffer, size_t size) override
{
std::string message((const char*)buffer, size);
std::cout << "Incoming: " << message << std::endl;
// Multicast message to all connected sessions
server()->Multicast(message);
// If the buffer starts with '!' the disconnect the current session
if (message == "!")
DisconnectAsync();
}
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "Chat SSL session caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
};
class ChatServer : public CppServer::Asio::SSLServer
{
public:
using CppServer::Asio::SSLServer::SSLServer;
protected:
std::shared_ptr<CppServer::Asio::SSLSession> CreateSession(std::shared_ptr<CppServer::Asio::SSLServer> server) override
{
return std::make_shared<ChatSession>(server);
}
protected:
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "Chat TCP server caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
};
int main(int argc, char** argv)
{
// SSL server port
int port = 2222;
if (argc > 1)
port = std::atoi(argv[1]);
std::cout << "SSL server port: " << port << std::endl;
// Create a new Asio service
auto service = std::make_shared<CppServer::Asio::Service>();
// Start the Asio service
std::cout << "Asio service starting...";
service->Start();
std::cout << "Done!" << std::endl;
// Create and prepare a new SSL server context
auto context = std::make_shared<CppServer::Asio::SSLContext>(asio::ssl::context::tlsv12);
context->set_password_callback([](size_t max_length, asio::ssl::context::password_purpose purpose) -> std::string { return "qwerty"; });
context->use_certificate_chain_file("../tools/certificates/server.pem");
context->use_private_key_file("../tools/certificates/server.pem", asio::ssl::context::pem);
context->use_tmp_dh_file("../tools/certificates/dh4096.pem");
// Create a new SSL chat server
auto server = std::make_shared<ChatServer>(service, context, port);
// Start the server
std::cout << "Server starting...";
server->Start();
std::cout << "Done!" << std::endl;
std::cout << "Press Enter to stop the server or '!' to restart the server..." << std::endl;
// Perform text input
std::string line;
while (getline(std::cin, line))
{
if (line.empty())
break;
// Restart the server
if (line == "!")
{
std::cout << "Server restarting...";
server->Restart();
std::cout << "Done!" << std::endl;
continue;
}
// Multicast admin message to all sessions
line = "(admin) " + line;
server->Multicast(line);
}
// Stop the server
std::cout << "Server stopping...";
server->Stop();
std::cout << "Done!" << std::endl;
// Stop the Asio service
std::cout << "Asio service stopping...";
service->Stop();
std::cout << "Done!" << std::endl;
return 0;
}
Example: SSL chat client
Here comes the example of the SSL chat client. It connects to the SSL chat server and allows to send message to it and receive new messages.
This example is very similar to the TCP one except the code that prepares SSL context and handshake handler.
#include "server/asio/ssl_client.h"
#include "threads/thread.h"
#include <atomic>
#include <iostream>
class ChatClient : public CppServer::Asio::SSLClient
{
public:
ChatClient(std::shared_ptr<CppServer::Asio::Service> service, std::shared_ptr<CppServer::Asio::SSLContext> context, const std::string& address, int port)
: CppServer::Asio::SSLClient(service, context, address, port)
{
_stop = false;
}
void DisconnectAndStop()
{
_stop = true;
DisconnectAsync();
while (IsConnected())
CppCommon::Thread::Yield();
}
protected:
void onConnected() override
{
std::cout << "Chat SSL client connected a new session with Id " << id() << std::endl;
}
void onHandshaked() override
{
std::cout << "Chat SSL client handshaked a new session with Id " << id() << std::endl;
}
void onDisconnected() override
{
std::cout << "Chat SSL client disconnected a session with Id " << id() << std::endl;
// Wait for a while...
CppCommon::Thread::Sleep(1000);
// Try to connect again
if (!_stop)
ConnectAsync();
}
void onReceived(const void* buffer, size_t size) override
{
std::cout << "Incoming: " << std::string((const char*)buffer, size) << std::endl;
}
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "Chat SSL client caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
private:
std::atomic<bool> _stop;
};
int main(int argc, char** argv)
{
// SSL server address
std::string address = "127.0.0.1";
if (argc > 1)
address = argv[1];
// SSL server port
int port = 2222;
if (argc > 2)
port = std::atoi(argv[2]);
std::cout << "SSL server address: " << address << std::endl;
std::cout << "SSL server port: " << port << std::endl;
// Create a new Asio service
auto service = std::make_shared<CppServer::Asio::Service>();
// Start the Asio service
std::cout << "Asio service starting...";
service->Start();
std::cout << "Done!" << std::endl;
// Create and prepare a new SSL client context
auto context = std::make_shared<CppServer::Asio::SSLContext>(asio::ssl::context::tlsv12);
context->set_default_verify_paths();
context->set_root_certs();
context->set_verify_mode(asio::ssl::verify_peer | asio::ssl::verify_fail_if_no_peer_cert);
context->load_verify_file("../tools/certificates/ca.pem");
// Create a new SSL chat client
auto client = std::make_shared<ChatClient>(service, context, address, port);
// Connect the client
std::cout << "Client connecting...";
client->ConnectAsync();
std::cout << "Done!" << std::endl;
std::cout << "Press Enter to stop the client or '!' to reconnect the client..." << std::endl;
// Perform text input
std::string line;
while (getline(std::cin, line))
{
if (line.empty())
break;
// Disconnect the client
if (line == "!")
{
std::cout << "Client disconnecting...";
client->DisconnectAsync();
std::cout << "Done!" << std::endl;
continue;
}
// Send the entered text to the chat server
client->SendAsync(line);
}
// Disconnect the client
std::cout << "Client disconnecting...";
client->DisconnectAndStop();
std::cout << "Done!" << std::endl;
// Stop the Asio service
std::cout << "Asio service stopping...";
service->Stop();
std::cout << "Done!" << std::endl;
return 0;
}
Example: UDP echo server
Here comes the example of the UDP echo server. It receives a datagram mesage from any UDP client and resend it back without any changes.
#include "server/asio/udp_server.h"
#include "threads/thread.h"
#include <iostream>
class EchoServer : public CppServer::Asio::UDPServer
{
public:
using CppServer::Asio::UDPServer::UDPServer;
protected:
void onStarted() override
{
// Start receive datagrams
ReceiveAsync();
}
void onReceived(const asio::ip::udp::endpoint& endpoint, const void* buffer, size_t size) override
{
std::string message((const char*)buffer, size);
std::cout << "Incoming: " << message << std::endl;
// Echo the message back to the sender
SendAsync(endpoint, message);
}
void onSent(const asio::ip::udp::endpoint& endpoint, size_t sent) override
{
// Continue receive datagrams
ReceiveAsync();
}
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "Echo UDP server caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
};
int main(int argc, char** argv)
{
// UDP server port
int port = 3333;
if (argc > 1)
port = std::atoi(argv[1]);
std::cout << "UDP server port: " << port << std::endl;
// Create a new Asio service
auto service = std::make_shared<CppServer::Asio::Service>();
// Start the Asio service
std::cout << "Asio service starting...";
service->Start();
std::cout << "Done!" << std::endl;
// Create a new UDP echo server
auto server = std::make_shared<EchoServer>(service, port);
// Start the server
std::cout << "Server starting...";
server->Start();
std::cout << "Done!" << std::endl;
std::cout << "Press Enter to stop the server or '!' to restart the server..." << std::endl;
// Perform text input
std::string line;
while (getline(std::cin, line))
{
if (line.empty())
break;
// Restart the server
if (line == "!")
{
std::cout << "Server restarting...";
server->Restart();
std::cout << "Done!" << std::endl;
continue;
}
}
// Stop the server
std::cout << "Server stopping...";
server->Stop();
std::cout << "Done!" << std::endl;
// Stop the Asio service
std::cout << "Asio service stopping...";
service->Stop();
std::cout << "Done!" << std::endl;
return 0;
}
Example: UDP echo client
Here comes the example of the UDP echo client. It sends user datagram message to UDP server and listen for response.
#include "server/asio/udp_client.h"
#include "threads/thread.h"
#include <atomic>
#include <iostream>
class EchoClient : public CppServer::Asio::UDPClient
{
public:
EchoClient(std::shared_ptr<CppServer::Asio::Service> service, const std::string& address, int port)
: CppServer::Asio::UDPClient(service, address, port)
{
_stop = false;
}
void DisconnectAndStop()
{
_stop = true;
DisconnectAsync();
while (IsConnected())
CppCommon::Thread::Yield();
}
protected:
void onConnected() override
{
std::cout << "Echo UDP client connected a new session with Id " << id() << std::endl;
// Start receive datagrams
ReceiveAsync();
}
void onDisconnected() override
{
std::cout << "Echo UDP client disconnected a session with Id " << id() << std::endl;
// Wait for a while...
CppCommon::Thread::Sleep(1000);
// Try to connect again
if (!_stop)
ConnectAsync();
}
void onReceived(const asio::ip::udp::endpoint& endpoint, const void* buffer, size_t size) override
{
std::cout << "Incoming: " << std::string((const char*)buffer, size) << std::endl;
// Continue receive datagrams
ReceiveAsync();
}
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "Echo UDP client caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
private:
std::atomic<bool> _stop;
};
int main(int argc, char** argv)
{
// UDP server address
std::string address = "127.0.0.1";
if (argc > 1)
address = argv[1];
// UDP server port
int port = 3333;
if (argc > 2)
port = std::atoi(argv[2]);
std::cout << "UDP server address: " << address << std::endl;
std::cout << "UDP server port: " << port << std::endl;
// Create a new Asio service
auto service = std::make_shared<CppServer::Asio::Service>();
// Start the Asio service
std::cout << "Asio service starting...";
service->Start();
std::cout << "Done!" << std::endl;
// Create a new UDP echo client
auto client = std::make_shared<EchoClient>(service, address, port);
// Connect the client
std::cout << "Client connecting...";
client->ConnectAsync();
std::cout << "Done!" << std::endl;
std::cout << "Press Enter to stop the client or '!' to reconnect the client..." << std::endl;
// Perform text input
std::string line;
while (getline(std::cin, line))
{
if (line.empty())
break;
// Disconnect the client
if (line == "!")
{
std::cout << "Client disconnecting...";
client->DisconnectAsync();
std::cout << "Done!" << std::endl;
continue;
}
// Send the entered text to the echo server
client->SendSync(line);
}
// Disconnect the client
std::cout << "Client disconnecting...";
client->DisconnectAndStop();
std::cout << "Done!" << std::endl;
// Stop the Asio service
std::cout << "Asio service stopping...";
service->Stop();
std::cout << "Done!" << std::endl;
return 0;
}
Example: UDP multicast server
Here comes the example of the UDP multicast server. It use multicast IP address to multicast datagram messages to all client that joined corresponding UDP multicast group.
#include "server/asio/udp_server.h"
#include "threads/thread.h"
#include <iostream>
class MulticastServer : public CppServer::Asio::UDPServer
{
public:
using CppServer::Asio::UDPServer::UDPServer;
protected:
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "Multicast UDP server caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
};
int main(int argc, char** argv)
{
// UDP multicast address
std::string multicast_address = "239.255.0.1";
if (argc > 1)
multicast_address = argv[1];
// UDP multicast port
int multicast_port = 3334;
if (argc > 2)
multicast_port = std::atoi(argv[2]);
std::cout << "UDP multicast address: " << multicast_address << std::endl;
std::cout << "UDP multicast port: " << multicast_port << std::endl;
// Create a new Asio service
auto service = std::make_shared<CppServer::Asio::Service>();
// Start the Asio service
std::cout << "Asio service starting...";
service->Start();
std::cout << "Done!" << std::endl;
// Create a new UDP multicast server
auto server = std::make_shared<MulticastServer>(service, 0);
// Start the multicast server
std::cout << "Server starting...";
server->Start(multicast_address, multicast_port);
std::cout << "Done!" << std::endl;
std::cout << "Press Enter to stop the server or '!' to restart the server..." << std::endl;
// Perform text input
std::string line;
while (getline(std::cin, line))
{
if (line.empty())
break;
// Restart the server
if (line == "!")
{
std::cout << "Server restarting...";
server->Restart();
std::cout << "Done!" << std::endl;
continue;
}
// Multicast admin message to all sessions
line = "(admin) " + line;
server->MulticastSync(line);
}
// Stop the server
std::cout << "Server stopping...";
server->Stop();
std::cout << "Done!" << std::endl;
// Stop the Asio service
std::cout << "Asio service stopping...";
service->Stop();
std::cout << "Done!" << std::endl;
return 0;
}
Example: UDP multicast client
Here comes the example of the UDP multicast client. It use multicast IP address and joins UDP multicast group in order to receive multicasted datagram messages from UDP server.
#include "server/asio/udp_client.h"
#include "threads/thread.h"
#include <atomic>
#include <iostream>
class MulticastClient : public CppServer::Asio::UDPClient
{
public:
MulticastClient(std::shared_ptr<CppServer::Asio::Service> service, const std::string& address, const std::string& multicast, int port)
: CppServer::Asio::UDPClient(service, address, port),
_multicast(multicast)
{
_stop = false;
}
void DisconnectAndStop()
{
_stop = true;
DisconnectAsync();
while (IsConnected())
CppCommon::Thread::Yield();
}
protected:
void onConnected() override
{
std::cout << "Multicast UDP client connected a new session with Id " << id() << std::endl;
// Join UDP multicast group
JoinMulticastGroupAsync(_multicast);
// Start receive datagrams
ReceiveAsync();
}
void onDisconnected() override
{
std::cout << "Multicast UDP client disconnected a session with Id " << id() << std::endl;
// Wait for a while...
CppCommon::Thread::Sleep(1000);
// Try to connect again
if (!_stop)
ConnectAsync();
}
void onReceived(const asio::ip::udp::endpoint& endpoint, const void* buffer, size_t size) override
{
std::cout << "Incoming: " << std::string((const char*)buffer, size) << std::endl;
// Continue receive datagrams
ReceiveAsync();
}
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "Multicast UDP client caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
private:
std::atomic<bool> _stop;
std::string _multicast;
};
int main(int argc, char** argv)
{
// UDP listen address
std::string listen_address = "0.0.0.0";
if (argc > 1)
listen_address = argv[1];
// UDP multicast address
std::string multicast_address = "239.255.0.1";
if (argc > 2)
multicast_address = argv[2];
// UDP multicast port
int multicast_port = 3334;
if (argc > 3)
multicast_port = std::atoi(argv[3]);
std::cout << "UDP listen address: " << listen_address << std::endl;
std::cout << "UDP multicast address: " << multicast_address << std::endl;
std::cout << "UDP multicast port: " << multicast_port << std::endl;
// Create a new Asio service
auto service = std::make_shared<CppServer::Asio::Service>();
// Start the Asio service
std::cout << "Asio service starting...";
service->Start();
std::cout << "Done!" << std::endl;
// Create a new UDP multicast client
auto client = std::make_shared<MulticastClient>(service, listen_address, multicast_address, multicast_port);
client->SetupMulticast(true);
// Connect the client
std::cout << "Client connecting...";
client->ConnectAsync();
std::cout << "Done!" << std::endl;
std::cout << "Press Enter to stop the client or '!' to reconnect the client..." << std::endl;
// Perform text input
std::string line;
while (getline(std::cin, line))
{
if (line.empty())
break;
// Disconnect the client
if (line == "!")
{
std::cout << "Client disconnecting...";
client->DisconnectAsync();
std::cout << "Done!" << std::endl;
continue;
}
}
// Disconnect the client
std::cout << "Client disconnecting...";
client->DisconnectAndStop();
std::cout << "Done!" << std::endl;
// Stop the Asio service
std::cout << "Asio service stopping...";
service->Stop();
std::cout << "Done!" << std::endl;
return 0;
}
Example: Simple protocol
Simple protocol is defined in simple.fbe file:
/*
Simple Fast Binary Encoding protocol for CppServer
https://github.com/chronoxor/FastBinaryEncoding
Generate protocol command: fbec --cpp --proto --input=simple.fbe --output=.
*/
// Domain declaration
domain com.chronoxor
// Package declaration
package simple
// Protocol version
version 1.0
// Simple request message
[request]
[response(SimpleResponse)]
[reject(SimpleReject)]
message SimpleRequest
{
// Request Id
uuid [id] = uuid1;
// Request message
string Message;
}
// Simple response
message SimpleResponse
{
// Response Id
uuid [id] = uuid1;
// Calculated message hash
uint32 Hash;
}
// Simple reject
message SimpleReject
{
// Reject Id
uuid [id] = uuid1;
// Error message
string Error;
}
// Simple notification
message SimpleNotify
{
// Server notification
string Notification;
}
// Disconnect request message
[request]
message DisconnectRequest
{
// Request Id
uuid [id] = uuid1;
}
Example: Simple protocol server
Here comes the example of the simple protocol server. It process client requests, answer with corresponding responses and send server notifications back to clients.
#include "asio_service.h"
#include "server/asio/tcp_server.h"
#include "../proto/simple_protocol.h"
#include <iostream>
class SimpleProtoSession : public CppServer::Asio::TCPSession, public FBE::simple::Sender, public FBE::simple::Receiver
{
public:
using CppServer::Asio::TCPSession::TCPSession;
protected:
void onConnected() override
{
std::cout << "Simple protocol session with Id " << id() << " connected!" << std::endl;
// Send invite notification
simple::SimpleNotify notify;
notify.Notification = "Hello from Simple protocol server! Please send a message or '!' to disconnect the client!";
send(notify);
}
void onDisconnected() override
{
std::cout << "Simple protocol session with Id " << id() << " disconnected!" << std::endl;
}
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "Simple protocol session caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
// Protocol handlers
void onReceive(const ::simple::DisconnectRequest& request) override { Disconnect(); }
void onReceive(const ::simple::SimpleRequest& request) override
{
std::cout << "Received: " << request << std::endl;
// Validate request
if (request.Message.empty())
{
// Send reject
simple::SimpleReject reject;
reject.id = request.id;
reject.Error = "Request message is empty!";
send(reject);
return;
}
static std::hash<std::string> hasher;
// Send response
simple::SimpleResponse response;
response.id = request.id;
response.Hash = (uint32_t)hasher(request.Message);
send(response);
}
// Protocol implementation
void onReceived(const void* buffer, size_t size) override { receive(buffer, size); }
size_t onSend(const void* data, size_t size) override { return SendAsync(data, size) ? size : 0; }
};
class SimpleProtoServer : public CppServer::Asio::TCPServer, public FBE::simple::Sender
{
public:
using CppServer::Asio::TCPServer::TCPServer;
protected:
std::shared_ptr<CppServer::Asio::TCPSession> CreateSession(const std::shared_ptr<CppServer::Asio::TCPServer>& server) override
{
return std::make_shared<SimpleProtoSession>(server);
}
protected:
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "Simple protocol server caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
// Protocol implementation
size_t onSend(const void* data, size_t size) override { Multicast(data, size); return size; }
};
int main(int argc, char** argv)
{
// Simple protocol server port
int port = 4444;
if (argc > 1)
port = std::atoi(argv[1]);
std::cout << "Simple protocol server port: " << port << std::endl;
std::cout << std::endl;
// Create a new Asio service
auto service = std::make_shared<AsioService>();
// Start the Asio service
std::cout << "Asio service starting...";
service->Start();
std::cout << "Done!" << std::endl;
// Create a new simple protocol server
auto server = std::make_shared<SimpleProtoServer>(service, port);
// Start the server
std::cout << "Server starting...";
server->Start();
std::cout << "Done!" << std::endl;
std::cout << "Press Enter to stop the server or '!' to restart the server..." << std::endl;
// Perform text input
std::string line;
while (getline(std::cin, line))
{
if (line.empty())
break;
// Restart the server
if (line == "!")
{
std::cout << "Server restarting...";
server->Restart();
std::cout << "Done!" << std::endl;
continue;
}
// Multicast admin notification to all sessions
simple::SimpleNotify notify;
notify.Notification = "(admin) " + line;
server->send(notify);
}
// Stop the server
std::cout << "Server stopping...";
server->Stop();
std::cout << "Done!" << std::endl;
// Stop the Asio service
std::cout << "Asio service stopping...";
service->Stop();
std::cout << "Done!" << std::endl;
return 0;
}
Example: Simple protocol client
Here comes the example of the simple protocol client. It connects to the simple protocol server and allows to send requests to it and receive corresponding responses.
#include "asio_service.h"
#include "server/asio/tcp_client.h"
#include "threads/thread.h"
#include "../proto/simple_protocol.h"
#include <atomic>
#include <iostream>
class SimpleProtoClient : public CppServer::Asio::TCPClient, public FBE::simple::Client
{
public:
using CppServer::Asio::TCPClient::TCPClient;
void DisconnectAndStop()
{
_stop = true;
DisconnectAsync();
while (IsConnected())
CppCommon::Thread::Yield();
}
protected:
void onConnected() override
{
std::cout << "Simple protocol client connected a new session with Id " << id() << std::endl;
// Reset FBE protocol buffers
reset();
}
void onDisconnected() override
{
std::cout << "Simple protocol client disconnected a session with Id " << id() << std::endl;
// Wait for a while...
CppCommon::Thread::Sleep(1000);
// Try to connect again
if (!_stop)
ConnectAsync();
}
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "Simple protocol client caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
// Protocol handlers
void onReceive(const ::simple::DisconnectRequest& request) override { Client::onReceive(request); std::cout << "Received: " << request << std::endl; DisconnectAsync(); }
void onReceive(const ::simple::SimpleResponse& response) override { Client::onReceive(response); std::cout << "Received: " << response << std::endl; }
void onReceive(const ::simple::SimpleReject& reject) override { Client::onReceive(reject); std::cout << "Received: " << reject << std::endl; }
void onReceive(const ::simple::SimpleNotify& notify) override { Client::onReceive(notify); std::cout << "Received: " << notify << std::endl; }
// Protocol implementation
void onReceived(const void* buffer, size_t size) override { receive(buffer, size); }
size_t onSend(const void* data, size_t size) override { return SendAsync(data, size) ? size : 0; }
private:
std::atomic<bool> _stop{false};
};
int main(int argc, char** argv)
{
// TCP server address
std::string address = "127.0.0.1";
if (argc > 1)
address = argv[1];
// Simple protocol server port
int port = 4444;
if (argc > 2)
port = std::atoi(argv[2]);
std::cout << "Simple protocol server address: " << address << std::endl;
std::cout << "Simple protocol server port: " << port << std::endl;
std::cout << std::endl;
// Create a new Asio service
auto service = std::make_shared<AsioService>();
// Start the Asio service
std::cout << "Asio service starting...";
service->Start();
std::cout << "Done!" << std::endl;
// Create a new simple protocol client
auto client = std::make_shared<SimpleProtoClient>(service, address, port);
// Connect the client
std::cout << "Client connecting...";
client->ConnectAsync();
std::cout << "Done!" << std::endl;
std::cout << "Press Enter to stop the client or '!' to reconnect the client..." << std::endl;
// Perform text input
std::string line;
while (getline(std::cin, line))
{
if (line.empty())
break;
// Reconnect the client
if (line == "!")
{
std::cout << "Client reconnecting...";
client->IsConnected() ? client->ReconnectAsync() : client->ConnectAsync();
std::cout << "Done!" << std::endl;
continue;
}
// Send request to the simple protocol server
simple::SimpleRequest request;
request.Message = line;
auto response = client->request(request).get();
// Show string hash calculation result
std::cout << "Hash of '" << line << "' = " << std::format("0x{:8X}", response.Hash) << std::endl;
}
// Disconnect the client
std::cout << "Client disconnecting...";
client->DisconnectAndStop();
std::cout << "Done!" << std::endl;
// Stop the Asio service
std::cout << "Asio service stopping...";
service->Stop();
std::cout << "Done!" << std::endl;
return 0;
}
Example: HTTP server
Here comes the example of the HTTP cache server. It allows to manipulate cache data with HTTP methods (GET, POST, PUT and DELETE).
Use the following link to open Swagger OpenAPI iterative documentation: http://localhost:8080/api/index.html
#include "server/http/http_server.h"
#include "string/string_utils.h"
#include "utility/singleton.h"
#include <iostream>
#include <map>
#include <mutex>
class Cache : public CppCommon::Singleton<Cache>
{
friend CppCommon::Singleton<Cache>;
public:
std::string GetAllCache()
{
std::scoped_lock locker(_cache_lock);
std::string result;
result += "[\n";
for (const auto& item : _cache)
{
result += " {\n";
result += " \"key\": \"" + item.first + "\",\n";
result += " \"value\": \"" + item.second + "\",\n";
result += " },\n";
}
result += "]\n";
return result;
}
bool GetCacheValue(std::string_view key, std::string& value)
{
std::scoped_lock locker(_cache_lock);
auto it = _cache.find(key);
if (it != _cache.end())
{
value = it->second;
return true;
}
else
return false;
}
void PutCacheValue(std::string_view key, std::string_view value)
{
std::scoped_lock locker(_cache_lock);
auto it = _cache.emplace(key, value);
if (!it.second)
it.first->second = value;
}
bool DeleteCacheValue(std::string_view key, std::string& value)
{
std::scoped_lock locker(_cache_lock);
auto it = _cache.find(key);
if (it != _cache.end())
{
value = it->second;
_cache.erase(it);
return true;
}
else
return false;
}
private:
std::mutex _cache_lock;
std::map<std::string, std::string, std::less<>> _cache;
};
class HTTPCacheSession : public CppServer::HTTP::HTTPSession
{
public:
using CppServer::HTTP::HTTPSession::HTTPSession;
protected:
void onReceivedRequest(const CppServer::HTTP::HTTPRequest& request) override
{
// Show HTTP request content
std::cout << std::endl << request;
// Process HTTP request methods
if (request.method() == "HEAD")
SendResponseAsync(response().MakeHeadResponse());
else if (request.method() == "GET")
{
std::string key(request.url());
std::string value;
// Decode the key value
key = CppCommon::Encoding::URLDecode(key);
CppCommon::StringUtils::ReplaceFirst(key, "/api/cache", "");
CppCommon::StringUtils::ReplaceFirst(key, "?key=", "");
if (key.empty())
{
// Response with all cache values
SendResponseAsync(response().MakeGetResponse(Cache::GetInstance().GetAllCache(), "application/json; charset=UTF-8"));
}
// Get the cache value by the given key
else if (Cache::GetInstance().GetCacheValue(key, value))
{
// Response with the cache value
SendResponseAsync(response().MakeGetResponse(value));
}
else
SendResponseAsync(response().MakeErrorResponse(404, "Required cache value was not found for the key: " + key));
}
else if ((request.method() == "POST") || (request.method() == "PUT"))
{
std::string key(request.url());
std::string value(request.body());
// Decode the key value
key = CppCommon::Encoding::URLDecode(key);
CppCommon::StringUtils::ReplaceFirst(key, "/api/cache", "");
CppCommon::StringUtils::ReplaceFirst(key, "?key=", "");
// Put the cache value
Cache::GetInstance().PutCacheValue(key, value);
// Response with the cache value
SendResponseAsync(response().MakeOKResponse());
}
else if (request.method() == "DELETE")
{
std::string key(request.url());
std::string value;
// Decode the key value
key = CppCommon::Encoding::URLDecode(key);
CppCommon::StringUtils::ReplaceFirst(key, "/api/cache", "");
CppCommon::StringUtils::ReplaceFirst(key, "?key=", "");
// Delete the cache value
if (Cache::GetInstance().DeleteCacheValue(key, value))
{
// Response with the cache value
SendResponseAsync(response().MakeGetResponse(value));
}
else
SendResponseAsync(response().MakeErrorResponse(404, "Deleted cache value was not found for the key: " + key));
}
else if (request.method() == "OPTIONS")
SendResponseAsync(response().MakeOptionsResponse());
else if (request.method() == "TRACE")
SendResponseAsync(response().MakeTraceResponse(request.cache()));
else
SendResponseAsync(response().MakeErrorResponse("Unsupported HTTP method: " + std::string(request.method())));
}
void onReceivedRequestError(const CppServer::HTTP::HTTPRequest& request, const std::string& error) override
{
std::cout << "Request error: " << error << std::endl;
}
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "HTTP session caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
};
class HTTPCacheServer : public CppServer::HTTP::HTTPServer
{
public:
using CppServer::HTTP::HTTPServer::HTTPServer;
protected:
std::shared_ptr<CppServer::Asio::TCPSession> CreateSession(const std::shared_ptr<CppServer::Asio::TCPServer>& server) override
{
return std::make_shared<HTTPCacheSession>(std::dynamic_pointer_cast<CppServer::HTTP::HTTPServer>(server));
}
protected:
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "HTTP server caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
};
int main(int argc, char** argv)
{
// HTTP server port
int port = 8080;
if (argc > 1)
port = std::atoi(argv[1]);
// HTTP server content path
std::string www = "../www/api";
if (argc > 2)
www = argv[2];
std::cout << "HTTP server port: " << port << std::endl;
std::cout << "HTTP server static content path: " << www << std::endl;
std::cout << "HTTP server website: " << "http://localhost:" << port << "/api/index.html" << std::endl;
std::cout << std::endl;
// Create a new Asio service
auto service = std::make_shared<CppServer::Asio::Service>();
// Start the Asio service
std::cout << "Asio service starting...";
service->Start();
std::cout << "Done!" << std::endl;
// Create a new HTTP server
auto server = std::make_shared<HTTPCacheServer>(service, port);
server->AddStaticContent(www, "/api");
// Start the server
std::cout << "Server starting...";
server->Start();
std::cout << "Done!" << std::endl;
std::cout << "Press Enter to stop the server or '!' to restart the server..." << std::endl;
// Perform text input
std::string line;
while (getline(std::cin, line))
{
if (line.empty())
break;
// Restart the server
if (line == "!")
{
std::cout << "Server restarting...";
server->Restart();
std::cout << "Done!" << std::endl;
continue;
}
}
// Stop the server
std::cout << "Server stopping...";
server->Stop();
std::cout << "Done!" << std::endl;
// Stop the Asio service
std::cout << "Asio service stopping...";
service->Stop();
std::cout << "Done!" << std::endl;
return 0;
}
Example: HTTP client
Here comes the example of the HTTP client. It allows to send HTTP requests (GET, POST, PUT and DELETE) and receive HTTP responses.
#include "server/http/http_client.h"
#include "string/string_utils.h"
#include <iostream>
int main(int argc, char** argv)
{
// HTTP server address
std::string address = "127.0.0.1";
if (argc > 1)
address = argv[1];
std::cout << "HTTP server address: " << address << std::endl;
std::cout << std::endl;
// Create a new Asio service
auto service = std::make_shared<CppServer::Asio::Service>();
// Start the Asio service
std::cout << "Asio service starting...";
service->Start();
std::cout << "Done!" << std::endl;
// Create a new HTTP client
auto client = std::make_shared<CppServer::HTTP::HTTPClientEx>(service, address, "http");
std::cout << "Press Enter to stop the client or '!' to reconnect the client..." << std::endl;
try
{
// Perform text input
std::string line;
while (getline(std::cin, line))
{
if (line.empty())
break;
// Reconnect the client
if (line == "!")
{
std::cout << "Client reconnecting...";
client->ReconnectAsync();
std::cout << "Done!" << std::endl;
continue;
}
auto commands = CppCommon::StringUtils::Split(line, ' ', true);
if (commands.size() < 2)
{
std::cout << "HTTP method and URL must be entered!" << std::endl;
continue;
}
if (CppCommon::StringUtils::ToUpper(commands[0]) == "HEAD")
{
auto response = client->SendHeadRequest(commands[1]).get();
std::cout << response << std::endl;
}
else if (CppCommon::StringUtils::ToUpper(commands[0]) == "GET")
{
auto response = client->SendGetRequest(commands[1]).get();
std::cout << response << std::endl;
}
else if (CppCommon::StringUtils::ToUpper(commands[0]) == "POST")
{
if (commands.size() < 3)
{
std::cout << "HTTP method, URL and body must be entered!" << std::endl;
continue;
}
auto response = client->SendPostRequest(commands[1], commands[2]).get();
std::cout << response << std::endl;
}
else if (CppCommon::StringUtils::ToUpper(commands[0]) == "PUT")
{
if (commands.size() < 3)
{
std::cout << "HTTP method, URL and body must be entered!" << std::endl;
continue;
}
auto response = client->SendPutRequest(commands[1], commands[2]).get();
std::cout << response << std::endl;
}
else if (CppCommon::StringUtils::ToUpper(commands[0]) == "DELETE")
{
auto response = client->SendDeleteRequest(commands[1]).get();
std::cout << response << std::endl;
}
else if (CppCommon::StringUtils::ToUpper(commands[0]) == "OPTIONS")
{
auto response = client->SendOptionsRequest(commands[1]).get();
std::cout << response << std::endl;
}
else if (CppCommon::StringUtils::ToUpper(commands[0]) == "TRACE")
{
auto response = client->SendTraceRequest(commands[1]).get();
std::cout << response << std::endl;
}
else
std::cout << "Unknown HTTP method: " << commands[0] << std::endl;
}
}
catch (const std::exception& ex)
{
std::cerr << ex.what() << std::endl;
}
// Stop the Asio service
std::cout << "Asio service stopping...";
service->Stop();
std::cout << "Done!" << std::endl;
return 0;
}
Example: HTTPS server
Here comes the example of the HTTPS cache server. It allows to manipulate cache data with HTTP methods (GET, POST, PUT and DELETE) with secured transport protocol.
Use the following link to open Swagger OpenAPI iterative documentation: https://localhost:8443/api/index.html
#include "server/http/https_server.h"
#include "string/string_utils.h"
#include "utility/singleton.h"
#include <iostream>
#include <map>
#include <mutex>
class Cache : public CppCommon::Singleton<Cache>
{
friend CppCommon::Singleton<Cache>;
public:
std::string GetAllCache()
{
std::scoped_lock locker(_cache_lock);
std::string result;
result += "[\n";
for (const auto& item : _cache)
{
result += " {\n";
result += " \"key\": \"" + item.first + "\",\n";
result += " \"value\": \"" + item.second + "\",\n";
result += " },\n";
}
result += "]\n";
return result;
}
bool GetCacheValue(std::string_view key, std::string& value)
{
std::scoped_lock locker(_cache_lock);
auto it = _cache.find(key);
if (it != _cache.end())
{
value = it->second;
return true;
}
else
return false;
}
void PutCacheValue(std::string_view key, std::string_view value)
{
std::scoped_lock locker(_cache_lock);
auto it = _cache.emplace(key, value);
if (!it.second)
it.first->second = value;
}
bool DeleteCacheValue(std::string_view key, std::string& value)
{
std::scoped_lock locker(_cache_lock);
auto it = _cache.find(key);
if (it != _cache.end())
{
value = it->second;
_cache.erase(it);
return true;
}
else
return false;
}
private:
std::mutex _cache_lock;
std::map<std::string, std::string, std::less<>> _cache;
};
class HTTPSCacheSession : public CppServer::HTTP::HTTPSSession
{
public:
using CppServer::HTTP::HTTPSSession::HTTPSSession;
protected:
void onReceivedRequest(const CppServer::HTTP::HTTPRequest& request) override
{
// Show HTTP request content
std::cout << std::endl << request;
// Process HTTP request methods
if (request.method() == "HEAD")
SendResponseAsync(response().MakeHeadResponse());
else if (request.method() == "GET")
{
std::string key(request.url());
std::string value;
// Decode the key value
key = CppCommon::Encoding::URLDecode(key);
CppCommon::StringUtils::ReplaceFirst(key, "/api/cache", "");
CppCommon::StringUtils::ReplaceFirst(key, "?key=", "");
if (key.empty())
{
// Response with all cache values
SendResponseAsync(response().MakeGetResponse(Cache::GetInstance().GetAllCache(), "application/json; charset=UTF-8"));
}
// Get the cache value by the given key
else if (Cache::GetInstance().GetCacheValue(key, value))
{
// Response with the cache value
SendResponseAsync(response().MakeGetResponse(value));
}
else
SendResponseAsync(response().MakeErrorResponse(404, "Required cache value was not found for the key: " + key));
}
else if ((request.method() == "POST") || (request.method() == "PUT"))
{
std::string key(request.url());
std::string value(request.body());
// Decode the key value
key = CppCommon::Encoding::URLDecode(key);
CppCommon::StringUtils::ReplaceFirst(key, "/api/cache", "");
CppCommon::StringUtils::ReplaceFirst(key, "?key=", "");
// Put the cache value
Cache::GetInstance().PutCacheValue(key, value);
// Response with the cache value
SendResponseAsync(response().MakeOKResponse());
}
else if (request.method() == "DELETE")
{
std::string key(request.url());
std::string value;
// Decode the key value
key = CppCommon::Encoding::URLDecode(key);
CppCommon::StringUtils::ReplaceFirst(key, "/api/cache", "");
CppCommon::StringUtils::ReplaceFirst(key, "?key=", "");
// Delete the cache value
if (Cache::GetInstance().DeleteCacheValue(key, value))
{
// Response with the cache value
SendResponseAsync(response().MakeGetResponse(value));
}
else
SendResponseAsync(response().MakeErrorResponse(404, "Deleted cache value was not found for the key: " + key));
}
else if (request.method() == "OPTIONS")
SendResponseAsync(response().MakeOptionsResponse());
else if (request.method() == "TRACE")
SendResponseAsync(response().MakeTraceResponse(request.cache()));
else
SendResponseAsync(response().MakeErrorResponse("Unsupported HTTP method: " + std::string(request.method())));
}
void onReceivedRequestError(const CppServer::HTTP::HTTPRequest& request, const std::string& error) override
{
std::cout << "Request error: " << error << std::endl;
}
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "HTTPS session caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
};
class HTTPSCacheServer : public CppServer::HTTP::HTTPSServer
{
public:
using CppServer::HTTP::HTTPSServer::HTTPSServer;
protected:
std::shared_ptr<CppServer::Asio::SSLSession> CreateSession(const std::shared_ptr<CppServer::Asio::SSLServer>& server) override
{
return std::make_shared<HTTPSCacheSession>(std::dynamic_pointer_cast<CppServer::HTTP::HTTPSServer>(server));
}
protected:
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "HTTPS server caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
};
int main(int argc, char** argv)
{
// HTTPS server port
int port = 8443;
if (argc > 1)
port = std::atoi(argv[1]);
// HTTPS server content path
std::string www = "../www/api";
if (argc > 2)
www = argv[2];
std::cout << "HTTPS server port: " << port << std::endl;
std::cout << "HTTPS server static content path: " << www << std::endl;
std::cout << "HTTPS server website: " << "https://localhost:" << port << "/api/index.html" << std::endl;
std::cout << std::endl;
// Create a new Asio service
auto service = std::make_shared<CppServer::Asio::Service>();
// Start the Asio service
std::cout << "Asio service starting...";
service->Start();
std::cout << "Done!" << std::endl;
// Create and prepare a new SSL server context
auto context = std::make_shared<CppServer::Asio::SSLContext>(asio::ssl::context::tlsv12);
context->set_password_callback([](size_t max_length, asio::ssl::context::password_purpose purpose) -> std::string { return "qwerty"; });
context->use_certificate_chain_file("../tools/certificates/server.pem");
context->use_private_key_file("../tools/certificates/server.pem", asio::ssl::context::pem);
context->use_tmp_dh_file("../tools/certificates/dh4096.pem");
// Create a new HTTPS server
auto server = std::make_shared<HTTPSCacheServer>(service, context, port);
server->AddStaticContent(www, "/api");
// Start the server
std::cout << "Server starting...";
server->Start();
std::cout << "Done!" << std::endl;
std::cout << "Press Enter to stop the server or '!' to restart the server..." << std::endl;
// Perform text input
std::string line;
while (getline(std::cin, line))
{
if (line.empty())
break;
// Restart the server
if (line == "!")
{
std::cout << "Server restarting...";
server->Restart();
std::cout << "Done!" << std::endl;
continue;
}
}
// Stop the server
std::cout << "Server stopping...";
server->Stop();
std::cout << "Done!" << std::endl;
// Stop the Asio service
std::cout << "Asio service stopping...";
service->Stop();
std::cout << "Done!" << std::endl;
return 0;
}
Example: HTTPS client
Here comes the example of the HTTPS client. It allows to send HTTP requests (GET, POST, PUT and DELETE) and receive HTTP responses with secured transport protocol.
#include "server/http/https_client.h"
#include "string/string_utils.h"
#include <iostream>
int main(int argc, char** argv)
{
// HTTP server address
std::string address = "127.0.0.1";
if (argc > 1)
address = argv[1];
std::cout << "HTTPS server address: " << address << std::endl;
std::cout << std::endl;
// Create a new Asio service
auto service = std::make_shared<CppServer::Asio::Service>();
// Start the Asio service
std::cout << "Asio service starting...";
service->Start();
std::cout << "Done!" << std::endl;
// Create and prepare a new SSL client context
auto context = std::make_shared<CppServer::Asio::SSLContext>(asio::ssl::context::tlsv12);
context->set_default_verify_paths();
context->set_root_certs();
context->set_verify_mode(asio::ssl::verify_peer | asio::ssl::verify_fail_if_no_peer_cert);
context->load_verify_file("../tools/certificates/ca.pem");
// Create a new HTTP client
auto client = std::make_shared<CppServer::HTTP::HTTPSClientEx>(service, context, address, "https");
std::cout << "Press Enter to stop the client or '!' to reconnect the client..." << std::endl;
try
{
// Perform text input
std::string line;
while (getline(std::cin, line))
{
if (line.empty())
break;
// Reconnect the client
if (line == "!")
{
std::cout << "Client reconnecting...";
client->ReconnectAsync();
std::cout << "Done!" << std::endl;
continue;
}
auto commands = CppCommon::StringUtils::Split(line, ' ', true);
if (commands.size() < 2)
{
std::cout << "HTTP method and URL must be entered!" << std::endl;
continue;
}
if (CppCommon::StringUtils::ToUpper(commands[0]) == "HEAD")
{
auto response = client->SendHeadRequest(commands[1]).get();
std::cout << response << std::endl;
}
else if (CppCommon::StringUtils::ToUpper(commands[0]) == "GET")
{
auto response = client->SendGetRequest(commands[1]).get();
std::cout << response << std::endl;
}
else if (CppCommon::StringUtils::ToUpper(commands[0]) == "POST")
{
if (commands.size() < 3)
{
std::cout << "HTTP method, URL and body must be entered!" << std::endl;
continue;
}
auto response = client->SendPostRequest(commands[1], commands[2]).get();
std::cout << response << std::endl;
}
else if (CppCommon::StringUtils::ToUpper(commands[0]) == "PUT")
{
if (commands.size() < 3)
{
std::cout << "HTTP method, URL and body must be entered!" << std::endl;
continue;
}
auto response = client->SendPutRequest(commands[1], commands[2]).get();
std::cout << response << std::endl;
}
else if (CppCommon::StringUtils::ToUpper(commands[0]) == "DELETE")
{
auto response = client->SendDeleteRequest(commands[1]).get();
std::cout << response << std::endl;
}
else if (CppCommon::StringUtils::ToUpper(commands[0]) == "OPTIONS")
{
auto response = client->SendOptionsRequest(commands[1]).get();
std::cout << response << std::endl;
}
else if (CppCommon::StringUtils::ToUpper(commands[0]) == "TRACE")
{
auto response = client->SendTraceRequest(commands[1]).get();
std::cout << response << std::endl;
}
else
std::cout << "Unknown HTTP method: " << commands[0] << std::endl;
}
}
catch (const std::exception& ex)
{
std::cerr << ex.what() << std::endl;
}
// Stop the Asio service
std::cout << "Asio service stopping...";
service->Stop();
std::cout << "Done!" << std::endl;
return 0;
}
Example: WebSocket chat server
Here comes the example of the WebSocket chat server. It handles multiple WebSocket client sessions and multicast received message from any session to all ones. Also it is possible to send admin message directly from the server.
Use the following link to open WebSocket chat server example: http://localhost:8080/chat/index.html
#include "server/ws/ws_server.h"
#include <iostream>
class ChatSession : public CppServer::WS::WSSession
{
public:
using CppServer::WS::WSSession::WSSession;
protected:
void onWSConnected(const CppServer::HTTP::HTTPRequest& request) override
{
std::cout << "Chat WebSocket session with Id " << id() << " connected!" << std::endl;
// Send invite message
std::string message("Hello from WebSocket chat! Please send a message or '!' to disconnect the client!");
SendTextAsync(message);
}
void onWSDisconnected() override
{
std::cout << "Chat WebSocket session with Id " << id() << " disconnected!" << std::endl;
}
void onWSReceived(const void* buffer, size_t size) override
{
std::string message((const char*)buffer, size);
std::cout << "Incoming: " << message << std::endl;
// Multicast message to all connected sessions
std::dynamic_pointer_cast<CppServer::WS::WSServer>(server())->MulticastText(message);
// If the buffer starts with '!' the disconnect the current session
if (message == "!")
Close(1000);
}
void onWSPing(const void* buffer, size_t size) override
{
SendPongAsync(buffer, size);
}
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "Chat WebSocket session caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
};
class ChatServer : public CppServer::WS::WSServer
{
public:
using CppServer::WS::WSServer::WSServer;
protected:
std::shared_ptr<CppServer::Asio::TCPSession> CreateSession(std::shared_ptr<CppServer::Asio::TCPServer> server) override
{
return std::make_shared<ChatSession>(std::dynamic_pointer_cast<CppServer::WS::WSServer>(server));
}
protected:
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "Chat WebSocket server caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
};
int main(int argc, char** argv)
{
// WebSocket server port
int port = 8080;
if (argc > 1)
port = std::atoi(argv[1]);
// WebSocket server content path
std::string www = "../www/ws";
if (argc > 2)
www = argv[2];
std::cout << "WebSocket server port: " << port << std::endl;
std::cout << "WebSocket server static content path: " << www << std::endl;
std::cout << "WebSocket server website: " << "http://localhost:" << port << "/chat/index.html" << std::endl;
std::cout << std::endl;
// Create a new Asio service
auto service = std::make_shared<CppServer::Asio::Service>();
// Start the Asio service
std::cout << "Asio service starting...";
service->Start();
std::cout << "Done!" << std::endl;
// Create a new WebSocket chat server
auto server = std::make_shared<ChatServer>(service, port);
server->AddStaticContent(www, "/chat");
// Start the server
std::cout << "Server starting...";
server->Start();
std::cout << "Done!" << std::endl;
std::cout << "Press Enter to stop the server or '!' to restart the server..." << std::endl;
// Perform text input
std::string line;
while (getline(std::cin, line))
{
if (line.empty())
break;
// Restart the server
if (line == "!")
{
std::cout << "Server restarting...";
server->Restart();
std::cout << "Done!" << std::endl;
continue;
}
// Multicast admin message to all sessions
line = "(admin) " + line;
server->MulticastText(line);
}
// Stop the server
std::cout << "Server stopping...";
server->Stop();
std::cout << "Done!" << std::endl;
// Stop the Asio service
std::cout << "Asio service stopping...";
service->Stop();
std::cout << "Done!" << std::endl;
return 0;
}
Example: WebSocket chat client
Here comes the example of the WebSocket chat client. It connects to the WebSocket chat server and allows to send message to it and receive new messages.
#include "server/ws/ws_client.h"
#include "threads/thread.h"
#include <atomic>
#include <iostream>
class ChatClient : public CppServer::WS::WSClient
{
public:
using CppServer::WS::WSClient::WSClient;
void DisconnectAndStop()
{
_stop = true;
CloseAsync(1000);
while (IsConnected())
CppCommon::Thread::Yield();
}
protected:
void onWSConnecting(CppServer::HTTP::HTTPRequest& request) override
{
request.SetBegin("GET", "/");
request.SetHeader("Host", "localhost");
request.SetHeader("Origin", "http://localhost");
request.SetHeader("Upgrade", "websocket");
request.SetHeader("Connection", "Upgrade");
request.SetHeader("Sec-WebSocket-Key", CppCommon::Encoding::Base64Encode(ws_nonce()));
request.SetHeader("Sec-WebSocket-Protocol", "chat, superchat");
request.SetHeader("Sec-WebSocket-Version", "13");
}
void onWSConnected(const CppServer::HTTP::HTTPResponse& response) override
{
std::cout << "Chat WebSocket client connected a new session with Id " << id() << std::endl;
}
void onWSDisconnected() override
{
std::cout << "Chat WebSocket client disconnected a session with Id " << id() << std::endl;
}
void onWSReceived(const void* buffer, size_t size) override
{
std::cout << "Incoming: " << std::string((const char*)buffer, size) << std::endl;
}
void onWSPing(const void* buffer, size_t size) override
{
SendPongAsync(buffer, size);
}
void onDisconnected() override
{
WSClient::onDisconnected();
// Wait for a while...
CppCommon::Thread::Sleep(1000);
// Try to connect again
if (!_stop)
ConnectAsync();
}
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "Chat WebSocket client caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
private:
std::atomic<bool> _stop{false};
};
int main(int argc, char** argv)
{
// WebSocket server address
std::string address = "127.0.0.1";
if (argc > 1)
address = argv[1];
// WebSocket server port
int port = 8080;
if (argc > 2)
port = std::atoi(argv[2]);
std::cout << "WebSocket server address: " << address << std::endl;
std::cout << "WebSocket server port: " << port << std::endl;
std::cout << std::endl;
// Create a new Asio service
auto service = std::make_shared<CppServer::Asio::Service>();
// Start the Asio service
std::cout << "Asio service starting...";
service->Start();
std::cout << "Done!" << std::endl;
// Create a new WebSocket chat client
auto client = std::make_shared<ChatClient>(service, address, port);
// Connect the client
std::cout << "Client connecting...";
client->ConnectAsync();
std::cout << "Done!" << std::endl;
std::cout << "Press Enter to stop the client or '!' to reconnect the client..." << std::endl;
// Perform text input
std::string line;
while (getline(std::cin, line))
{
if (line.empty())
break;
// Reconnect the client
if (line == "!")
{
std::cout << "Client reconnecting...";
client->ReconnectAsync();
std::cout << "Done!" << std::endl;
continue;
}
// Send the entered text to the chat server
client->SendTextAsync(line);
}
// Disconnect the client
std::cout << "Client disconnecting...";
client->DisconnectAndStop();
std::cout << "Done!" << std::endl;
// Stop the Asio service
std::cout << "Asio service stopping...";
service->Stop();
std::cout << "Done!" << std::endl;
return 0;
}
Example: WebSocket secure chat server
Here comes the example of the WebSocket secure chat server. It handles multiple WebSocket secure client sessions and multicast received message from any session to all ones. Also it is possible to send admin message directly from the server.
This example is very similar to the WebSocket one except the code that prepares WebSocket secure context and handshake handler.
Use the following link to open WebSocket secure chat server example: https://localhost:8443/chat/index.html
#include "server/ws/wss_server.h"
#include <iostream>
class ChatSession : public CppServer::WS::WSSSession
{
public:
using CppServer::WS::WSSSession::WSSSession;
protected:
void onWSConnected(const CppServer::HTTP::HTTPRequest& request) override
{
std::cout << "Chat WebSocket secure session with Id " << id() << " connected!" << std::endl;
// Send invite message
std::string message("Hello from WebSocket secure chat! Please send a message or '!' to disconnect the client!");
SendTextAsync(message);
}
void onWSDisconnected() override
{
std::cout << "Chat WebSocket secure session with Id " << id() << " disconnected!" << std::endl;
}
void onWSReceived(const void* buffer, size_t size) override
{
std::string message((const char*)buffer, size);
std::cout << "Incoming: " << message << std::endl;
// Multicast message to all connected sessions
std::dynamic_pointer_cast<CppServer::WS::WSSServer>(server())->MulticastText(message);
// If the buffer starts with '!' the disconnect the current session
if (message == "!")
Close(1000);
}
void onWSPing(const void* buffer, size_t size) override
{
SendPongAsync(buffer, size);
}
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "Chat WebSocket secure session caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
};
class ChatServer : public CppServer::WS::WSSServer
{
public:
using CppServer::WS::WSSServer::WSSServer;
protected:
std::shared_ptr<CppServer::Asio::SSLSession> CreateSession(std::shared_ptr<CppServer::Asio::SSLServer> server) override
{
return std::make_shared<ChatSession>(std::dynamic_pointer_cast<CppServer::WS::WSSServer>(server));
}
protected:
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "Chat WebSocket secure server caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
};
int main(int argc, char** argv)
{
// WebSocket secure server port
int port = 8443;
if (argc > 1)
port = std::atoi(argv[1]);
// WebSocket secure server content path
std::string www = "../www/wss";
if (argc > 2)
www = argv[2];
std::cout << "WebSocket secure server port: " << port << std::endl;
std::cout << "WebSocket secure server static content path: " << www << std::endl;
std::cout << "WebSocket server website: " << "https://localhost:" << port << "/chat/index.html" << std::endl;
std::cout << std::endl;
// Create a new Asio service
auto service = std::make_shared<CppServer::Asio::Service>();
// Start the Asio service
std::cout << "Asio service starting...";
service->Start();
std::cout << "Done!" << std::endl;
// Create and prepare a new SSL server context
auto context = std::make_shared<CppServer::Asio::SSLContext>(asio::ssl::context::tlsv12);
context->set_password_callback([](size_t max_length, asio::ssl::context::password_purpose purpose) -> std::string { return "qwerty"; });
context->use_certificate_chain_file("../tools/certificates/server.pem");
context->use_private_key_file("../tools/certificates/server.pem", asio::ssl::context::pem);
context->use_tmp_dh_file("../tools/certificates/dh4096.pem");
// Create a new WebSocket secure chat server
auto server = std::make_shared<ChatServer>(service, context, port);
server->AddStaticContent(www, "/chat");
// Start the server
std::cout << "Server starting...";
server->Start();
std::cout << "Done!" << std::endl;
std::cout << "Press Enter to stop the server or '!' to restart the server..." << std::endl;
// Perform text input
std::string line;
while (getline(std::cin, line))
{
if (line.empty())
break;
// Restart the server
if (line == "!")
{
std::cout << "Server restarting...";
server->Restart();
std::cout << "Done!" << std::endl;
continue;
}
// Multicast admin message to all sessions
line = "(admin) " + line;
server->MulticastText(line);
}
// Stop the server
std::cout << "Server stopping...";
server->Stop();
std::cout << "Done!" << std::endl;
// Stop the Asio service
std::cout << "Asio service stopping...";
service->Stop();
std::cout << "Done!" << std::endl;
return 0;
}
Example: WebSocket secure chat client
Here comes the example of the WebSocket secure chat client. It connects to the WebSocket secure chat server and allows to send message to it and receive new messages.
This example is very similar to the WebSocket one except the code that prepares WebSocket secure context and handshake handler.
#include "server/ws/wss_client.h"
#include "threads/thread.h"
#include <atomic>
#include <iostream>
class ChatClient : public CppServer::WS::WSSClient
{
public:
using CppServer::WS::WSSClient::WSSClient;
void DisconnectAndStop()
{
_stop = true;
CloseAsync(1000);
while (IsConnected())
CppCommon::Thread::Yield();
}
protected:
void onWSConnecting(CppServer::HTTP::HTTPRequest& request) override
{
request.SetBegin("GET", "/");
request.SetHeader("Host", "localhost");
request.SetHeader("Origin", "https://localhost");
request.SetHeader("Upgrade", "websocket");
request.SetHeader("Connection", "Upgrade");
request.SetHeader("Sec-WebSocket-Key", CppCommon::Encoding::Base64Encode(ws_nonce()));
request.SetHeader("Sec-WebSocket-Protocol", "chat, superchat");
request.SetHeader("Sec-WebSocket-Version", "13");
}
void onWSConnected(const CppServer::HTTP::HTTPResponse& response) override
{
std::cout << "Chat WebSocket secure client connected a new session with Id " << id() << std::endl;
}
void onWSDisconnected() override
{
std::cout << "Chat WebSocket secure client disconnected a session with Id " << id() << std::endl;
}
void onWSReceived(const void* buffer, size_t size) override
{
std::cout << "Incoming: " << std::string((const char*)buffer, size) << std::endl;
}
void onWSPing(const void* buffer, size_t size) override
{
SendPongAsync(buffer, size);
}
void onDisconnected() override
{
WSSClient::onDisconnected();
// Wait for a while...
CppCommon::Thread::Sleep(1000);
// Try to connect again
if (!_stop)
ConnectAsync();
}
void onError(int error, const std::string& category, const std::string& message) override
{
std::cout << "Chat WebSocket secure client caught an error with code " << error << " and category '" << category << "': " << message << std::endl;
}
private:
std::atomic<bool> _stop{false};
};
int main(int argc, char** argv)
{
// WebSocket server address
std::string address = "127.0.0.1";
if (argc > 1)
address = argv[1];
// WebSocket server port
int port = 8443;
if (argc > 2)
port = std::atoi(argv[2]);
std::cout << "WebSocket secure server address: " << address << std::endl;
std::cout << "WebSocket secure server port: " << port << std::endl;
std::cout << std::endl;
// Create a new Asio service
auto service = std::make_shared<CppServer::Asio::Service>();
// Start the Asio service
std::cout << "Asio service starting...";
service->Start();
std::cout << "Done!" << std::endl;
// Create and prepare a new SSL client context
auto context = std::make_shared<CppServer::Asio::SSLContext>(asio::ssl::context::tlsv12);
context->set_default_verify_paths();
context->set_root_certs();
context->set_verify_mode(asio::ssl::verify_peer | asio::ssl::verify_fail_if_no_peer_cert);
context->load_verify_file("../tools/certificates/ca.pem");
// Create a new WebSocket chat client
auto client = std::make_shared<ChatClient>(service, context, address, port);
// Connect the client
std::cout << "Client connecting...";
client->ConnectAsync();
std::cout << "Done!" << std::endl;
std::cout << "Press Enter to stop the client or '!' to reconnect the client..." << std::endl;
// Perform text input
std::string line;
while (getline(std::cin, line))
{
if (line.empty())
break;
// Reconnect the client
if (line == "!")
{
std::cout << "Client reconnecting...";
client->ReconnectAsync();
std::cout << "Done!" << std::endl;
continue;
}
// Send the entered text to the chat server
client->SendTextAsync(line);
}
// Disconnect the client
std::cout << "Client disconnecting...";
client->DisconnectAndStop();
std::cout << "Done!" << std::endl;
// Stop the Asio service
std::cout << "Asio service stopping...";
service->Stop();
std::cout << "Done!" << std::endl;
return 0;
}
Performance
Here comes several communication scenarios with timing measurements.
Benchmark environment is the following:
CPU architecutre: Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz
CPU logical cores: 8
CPU physical cores: 4
CPU clock speed: 3.998 GHz
CPU Hyper-Threading: enabled
RAM total: 31.962 GiB
RAM free: 21.623 GiB
OS version: Microsoft Windows 8 Enterprise Edition (build 9200), 64-bit
OS bits: 64-bit
Process bits: 64-bit
Process configuaraion: release
Benchmark: Round-Trip
This scenario sends lots of messages from several clients to a server. The server responses to each message and resend the similar response to the client. The benchmark measures total round-trip time to send all messages and receive all responses, messages & data throughput, count of errors.
TCP echo server
- cppserver-performance-tcp_echo_server
- cppserver-performance-tcp_echo_client --clients 1 --threads 1
Server address: 127.0.0.1
Server port: 1111
Working threads: 1
Working clients: 1
Working messages: 1000
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 10.001 s
Total data: 1.692 GiB
Total messages: 56261685
Data throughput: 171.693 MiB/s
Message latency: 177 ns
Message throughput: 5625528 msg/s
- cppserver-performance-tcp_echo_server
- cppserver-performance-tcp_echo_client --clients 100 --threads 4
Server address: 127.0.0.1
Server port: 1111
Working threads: 4
Working clients: 100
Working messages: 1000
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 10.007 s
Total data: 1.151 GiB
Total messages: 38503396
Data throughput: 117.423 MiB/s
Message latency: 259 ns
Message throughput: 3847402 msg/s
SSL echo server
- cppserver-performance-ssl_echo_server
- cppserver-performance-ssl_echo_client --clients 1 --threads 1
Server address: 127.0.0.1
Server port: 2222
Working threads: 1
Working clients: 1
Working messages: 1000
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 10.012 s
Total data: 296.350 MiB
Total messages: 9710535
Data throughput: 29.612 MiB/s
Message latency: 1.031 mcs
Message throughput: 969878 msg/s
- cppserver-performance-ssl_echo_server
- cppserver-performance-ssl_echo_client --clients 100 --threads 4
Server address: 127.0.0.1
Server port: 2222
Working threads: 4
Working clients: 100
Working messages: 1000
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 10.341 s
Total data: 390.660 MiB
Total messages: 12800660
Data throughput: 37.792 MiB/s
Message latency: 807 ns
Message throughput: 1237782 msg/s
UDP echo server
- cppserver-performance-udp_echo_server
- cppserver-performance-udp_echo_client --clients 1 --threads 1
Server address: 127.0.0.1
Server port: 3333
Working threads: 1
Working clients: 1
Working messages: 1000
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 10.002 s
Total data: 46.032 MiB
Total messages: 1508355
Data throughput: 4.616 MiB/s
Message latency: 6.631 mcs
Message throughput: 150801 msg/s
- cppserver-performance-udp_echo_server
- cppserver-performance-udp_echo_client --clients 100 --threads 4
Server address: 127.0.0.1
Server port: 3333
Working threads: 4
Working clients: 100
Working messages: 1000
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 10.152 s
Total data: 32.185 MiB
Total messages: 1054512
Data throughput: 3.173 MiB/s
Message latency: 9.627 mcs
Message throughput: 103867 msg/s
Simple protocol server
- cppserver-performance-proto_server
- cppserver-performance-proto_client --clients 1 --threads 1
Server address: 127.0.0.1
Server port: 4444
Working threads: 1
Working clients: 1
Working messages: 1000
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 10.002 s
Total data: 497.096 MiB
Total messages: 16288783
Data throughput: 49.715 MiB/s
Message latency: 614 ns
Message throughput: 1628542 msg/s
- cppserver-performance-proto_server
- cppserver-performance-proto_client --clients 100 --threads 4
Server address: 127.0.0.1
Server port: 4444
Working threads: 4
Working clients: 100
Working messages: 1000
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 10.066 s
Total data: 997.384 MiB
Total messages: 32681995
Data throughput: 99.078 MiB/s
Message latency: 308 ns
Message throughput: 3246558 msg/s
WebSocket echo server
- cppserver-performance-ws_echo_server
- cppserver-performance-ws_echo_client --clients 1 --threads 1
Server address: 127.0.0.1
Server port: 8080
Working threads: 1
Working clients: 1
Working messages: 1000
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 9.994 s
Total data: 48.958 MiB
Total messages: 1603548
Data throughput: 4.918 MiB/s
Message latency: 6.232 mcs
Message throughput: 160448 msg/s
- cppserver-performance-ws_echo_server
- cppserver-performance-ws_echo_client --clients 100 --threads 4
Server address: 127.0.0.1
Server port: 8080
Working threads: 4
Working clients: 100
Working messages: 1000
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 11.402 s
Total data: 206.827 MiB
Total messages: 6776702
Data throughput: 18.140 MiB/s
Message latency: 1.682 mcs
Message throughput: 594328 msg/s
WebSocket secure echo server
- cppserver-performance-wss_echo_server
- cppserver-performance-wss_echo_client --clients 1 --threads 1
Server address: 127.0.0.1
Server port: 8443
Working threads: 1
Working clients: 1
Working messages: 1000
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 10.001 s
Total data: 62.068 MiB
Total messages: 2033811
Data throughput: 6.210 MiB/s
Message latency: 4.917 mcs
Message throughput: 203343 msg/s
- cppserver-performance-wss_echo_server
- cppserver-performance-wss_echo_client --clients 100 --threads 4
Server address: 127.0.0.1
Server port: 8443
Working threads: 4
Working clients: 100
Working messages: 1000
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 10.011 s
Total data: 249.1023 MiB
Total messages: 8191971
Data throughput: 24.993 MiB/s
Message latency: 1.222 mcs
Message throughput: 818230 msg/s
Benchmark: Multicast
In this scenario server multicasts messages to all connected clients. The benchmark counts total messages received by all clients for all the working time and measures messages & data throughput, count of errors.
TCP multicast server
- cppserver-performance-tcp_multicast_server
- cppserver-performance-tcp_multicast_client --clients 1 --threads 1
Server address: 127.0.0.1
Server port: 1111
Working threads: 1
Working clients: 1
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 10.001 s
Total data: 1.907 GiB
Total messages: 63283367
Data throughput: 193.103 MiB/s
Message latency: 158 ns
Message throughput: 6327549 msg/s
- cppserver-performance-tcp_multicast_server
- cppserver-performance-tcp_multicast_client --clients 100 --threads 4
Server address: 127.0.0.1
Server port: 1111
Working threads: 4
Working clients: 100
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 10.006 s
Total data: 1.1006 GiB
Total messages: 66535013
Data throughput: 202.930 MiB/s
Message latency: 150 ns
Message throughput: 6648899 msg/s
SSL multicast server
- cppserver-performance-ssl_multicast_server
- cppserver-performance-ssl_multicast_client --clients 1 --threads 1
Server address: 127.0.0.1
Server port: 2222
Working threads: 1
Working clients: 1
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 10.014 s
Total data: 1.535 GiB
Total messages: 51100073
Data throughput: 155.738 MiB/s
Message latency: 195 ns
Message throughput: 5102683 msg/s
- cppserver-performance-ssl_multicast_server
- cppserver-performance-ssl_multicast_client --clients 100 --threads 4
Server address: 127.0.0.1
Server port: 2222
Working threads: 4
Working clients: 100
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 10.691 s
Total data: 1.878 GiB
Total messages: 62334478
Data throughput: 177.954 MiB/s
Message latency: 171 ns
Message throughput: 5830473 msg/s
UDP multicast server
- cppserver-performance-udp_multicast_server
- cppserver-performance-udp_multicast_client --clients 1 --threads 1
Server address: 239.255.0.1
Server port: 3333
Working threads: 1
Working clients: 1
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 10.002 s
Total data: 23.777 MiB
Total messages: 778555
Data throughput: 2.384 MiB/s
Message latency: 12.847 mcs
Message throughput: 77833 msg/s
- cppserver-performance-udp_multicast_server
- cppserver-performance-udp_multicast_client --clients 100 --threads 4
Server address: 239.255.0.1
Server port: 3333
Working threads: 4
Working clients: 100
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 10.004 s
Total data: 52.457 MiB
Total messages: 1718575
Data throughput: 5.248 MiB/s
Message latency: 5.821 mcs
Message throughput: 171784 msg/s
WebSocket multicast server
- cppserver-performance-ws_multicast_server
- cppserver-performance-ws_multicast_client --clients 1 --threads 1
Server address: 127.0.0.1
Server port: 8080
Working threads: 1
Working clients: 1
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 10.001 s
Total data: 960.902 MiB
Total messages: 31486166
Data throughput: 96.075 MiB/s
Message latency: 317 ns
Message throughput: 3148135 msg/s
- cppserver-performance-ws_multicast_server
- cppserver-performance-ws_multicast_client --clients 100 --threads 4
Server address: 127.0.0.1
Server port: 8080
Working threads: 4
Working clients: 100
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 10.020 s
Total data: 986.489 MiB
Total messages: 32324898
Data throughput: 98.459 MiB/s
Message latency: 309 ns
Message throughput: 3225965 msg/s
WebSocket secure multicast server
- cppserver-performance-wss_multicast_server
- cppserver-performance-wss_multicast_client --clients 1 --threads 1
Server address: 127.0.0.1
Server port: 8443
Working threads: 1
Working clients: 1
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 10.002 s
Total data: 1.041 GiB
Total messages: 34903186
Data throughput: 106.505 MiB/s
Message latency: 286 ns
Message throughput: 3489578 msg/s
- cppserver-performance-wss_multicast_server
- cppserver-performance-wss_multicast_client --clients 100 --threads 4
Server address: 127.0.0.1
Server port: 8443
Working threads: 4
Working clients: 100
Message size: 32
Seconds to benchmarking: 10
Errors: 0
Total time: 10.013 s
Total data: 1.569 GiB
Total messages: 52225588
Data throughput: 159.172 MiB/s
Message latency: 191 ns
Message throughput: 5215639 msg/s
Benchmark: Web Server
HTTP Trace server
- cppserver-performance-http_trace_server
- cppserver-performance-http_trace_client --clients 1 --threads 1
Server address: 127.0.0.1
Server port: 80
Working threads: 1
Working clients: 1
Working messages: 1
Seconds to benchmarking: 10
Errors: 0
Total time: 10.001 s
Total data: 58.476 MiB
Total messages: 578353
Data throughput: 5.865 MiB/s
Message latency: 17.293 mcs
Message throughput: 57825 msg/s
- cppserver-performance-http_trace_server
- cppserver-performance-http_trace_client --clients 100 --threads 4
Server address: 127.0.0.1
Server port: 80
Working threads: 4
Working clients: 100
Working messages: 1
Seconds to benchmarking: 10
Errors: 0
Total time: 10.006 s
Total data: 310.730 MiB
Total messages: 3073650
Data throughput: 31.051 MiB/s
Message latency: 3.255 mcs
Message throughput: 307154 msg/s
HTTPS Trace server
- cppserver-performance-https_trace_server
- cppserver-performance-https_trace_client --clients 1 --threads 1
Server address: 127.0.0.1
Server port: 443
Working threads: 1
Working clients: 1
Working messages: 1
Seconds to benchmarking: 10
Errors: 0
Total time: 10.003 s
Total data: 37.475 MiB
Total messages: 370602
Data throughput: 3.763 MiB/s
Message latency: 26.992 mcs
Message throughput: 37047 msg/s
- cppserver-performance-https_trace_server
- cppserver-performance-https_trace_client --clients 100 --threads 4
Server address: 127.0.0.1
Server port: 443
Working threads: 4
Working clients: 100
Working messages: 1
Seconds to benchmarking: 10
Errors: 0
Total time: 10.035 s
Total data: 204.531 MiB
Total messages: 2023152
Data throughput: 20.389 MiB/s
Message latency: 4.960 mcs
Message throughput: 201602 msg/s
OpenSSL certificates
In order to create OpenSSL based server and client you should prepare a set of SSL certificates.
Production
Depending on your project, you may need to purchase a traditional SSL certificate signed by a Certificate Authority. If you, for instance, want some else's web browser to talk to your WebSocket project, you'll need a traditional SSL certificate.
Development
The commands below entered in the order they are listed will generate a self-signed certificate for development or testing purposes.
Certificate Authority
- Create CA private key
openssl genrsa -passout pass:qwerty -out ca-secret.key 4096- Remove passphrase
openssl rsa -passin pass:qwerty -in ca-secret.key -out ca.key- Create CA self-signed certificate
openssl req -new -x509 -days 3650 -subj '/C=BY/ST=Belarus/L=Minsk/O=Example root CA/OU=Example CA unit/CN=example.com' -key ca.key -out ca.crt- Convert CA self-signed certificate to PFX
openssl pkcs12 -export -passout pass:qwerty -inkey ca.key -in ca.crt -out ca.pfx- Convert CA self-signed certificate to PEM
openssl pkcs12 -passin pass:qwerty -passout pass:qwerty -in ca.pfx -out ca.pemSSL Server certificate
- Create private key for the server
openssl genrsa -passout pass:qwerty -out server-secret.key 4096- Remove passphrase
openssl rsa -passin pass:qwerty -in server-secret.key -out server.key- Create CSR for the server
openssl req -new -subj '/C=BY/ST=Belarus/L=Minsk/O=Example server/OU=Example server unit/CN=server.example.com' -key server.key -out server.csr- Create certificate for the server
openssl x509 -req -days 3650 -in server.csr -CA ca.crt -CAkey ca.key -set_serial 01 -out server.crt- Convert the server certificate to PFX
openssl pkcs12 -export -passout pass:qwerty -inkey server.key -in server.crt -out server.pfx- Convert the server certificate to PEM
openssl pkcs12 -passin pass:qwerty -passout pass:qwerty -in server.pfx -out server.pemSSL Client certificate
- Create private key for the client
openssl genrsa -passout pass:qwerty -out client-secret.key 4096- Remove passphrase
openssl rsa -passin pass:qwerty -in client-secret.key -out client.key- Create CSR for the client
openssl req -new -subj '/C=BY/ST=Belarus/L=Minsk/O=Example client/OU=Example client unit/CN=client.example.com' -key client.key -out client.csr- Create the client certificate
openssl x509 -req -days 3650 -in client.csr -CA ca.crt -CAkey ca.key -set_serial 01 -out client.crt- Convert the client certificate to PFX
openssl pkcs12 -export -passout pass:qwerty -inkey client.key -in client.crt -out client.pfx- Convert the client certificate to PEM
openssl pkcs12 -passin pass:qwerty -passout pass:qwerty -in client.pfx -out client.pemDiffie-Hellman key exchange
- Create DH parameters
openssl dhparam -out dh4096.pem 4096Author: chronoxor
Source Code: https://github.com/chronoxor/CppServer
License: MIT License
1649494800
Activity indicators for Modern C++
Indicators: Activity indicators for Modern C++
Highlights
- Thread-safe progress bars and spinners
- Header-only library. Grab a copy of
include/indicators. - Single-header version in
single_include/indicators. - Source for the above GIF can be found here
- MIT License
Basic Progress bar
To introduce a progress bar in your application, include indicators/progress_bar.hpp and create a ProgressBar object. Here's the general structure of a progress bar:
{prefix} {start} {fill} {lead} {remaining} {end} {percentage} [{elapsed}<{remaining}] {postfix}
^^^^^^^^^^^^^ Bar Width ^^^^^^^^^^^^^^^
The amount of progress in ProgressBar is maintained as a size_t in range [0, 100]. When progress reaches 100, the progression is complete.
From application-level code, there are two ways in which you can update this progress:
Update progress using bar.tick()
You can update the progress bar using bar.tick() which increments progress by exactly 1%.
#include <indicators/progress_bar.hpp>
#include <thread>
#include <chrono>
int main() {
using namespace indicators;
ProgressBar bar{
option::BarWidth{50},
option::Start{"["},
option::Fill{"="},
option::Lead{">"},
option::Remainder{" "},
option::End{"]"},
option::PostfixText{"Extracting Archive"},
option::ForegroundColor{Color::green},
option::FontStyles{std::vector<FontStyle>{FontStyle::bold}}
};
// Update bar state
while (true) {
bar.tick();
if (bar.is_completed())
break;
std::this_thread::sleep_for(std::chrono::milliseconds(100));
}
return 0;
}The above code will print a progress bar that goes from 0 to 100% at the rate of 1% every 100 ms.
Updating progress using bar.set_progress(value)
If you'd rather control progress of the bar in discrete steps, consider using bar.set_progress(value). Example:
#include <chrono>
#include <indicators/cursor_control.hpp>
#include <indicators/progress_bar.hpp>
#include <thread>
int main() {
using namespace indicators;
// Hide cursor
show_console_cursor(false);
ProgressBar bar{
option::BarWidth{50},
option::Start{"["},
option::Fill{"■"},
option::Lead{"■"},
option::Remainder{"-"},
option::End{" ]"},
option::PostfixText{"Loading dependency 1/4"},
option::ForegroundColor{Color::cyan},
option::FontStyles{std::vector<FontStyle>{FontStyle::bold}}
};
// Update bar state
bar.set_progress(10); // 10% done
// do some work
std::this_thread::sleep_for(std::chrono::milliseconds(800));
bar.set_option(option::PostfixText{"Loading dependency 2/4"});
bar.set_progress(30); // 30% done
// do some more work
std::this_thread::sleep_for(std::chrono::milliseconds(700));
bar.set_option(option::PostfixText{"Loading dependency 3/4"});
bar.set_progress(65); // 65% done
// do final bit of work
std::this_thread::sleep_for(std::chrono::milliseconds(900));
bar.set_option(option::PostfixText{"Loaded dependencies!"});
bar.set_progress(100); // all done
// Show cursor
show_console_cursor(true);
return 0;
}Showing Time Elapsed/Remaining
All progress bars and spinners in indicators support showing time elapsed and time remaining. Inspired by python's tqdm module, the format of this meter is [{elapsed}<{remaining}]:
#include <chrono>
#include <indicators/cursor_control.hpp>
#include <indicators/progress_bar.hpp>
#include <thread>
int main() {
using namespace indicators;
// Hide cursor
show_console_cursor(false);
indicators::ProgressBar bar{
option::BarWidth{50},
option::Start{" ["},
option::Fill{"█"},
option::Lead{"█"},
option::Remainder{"-"},
option::End{"]"},
option::PrefixText{"Training Gaze Network 👀"},
option::ForegroundColor{Color::yellow},
option::ShowElapsedTime{true},
option::ShowRemainingTime{true},
option::FontStyles{std::vector<FontStyle>{FontStyle::bold}}
};
// Update bar state
while (true) {
bar.tick();
if (bar.is_completed())
break;
std::this_thread::sleep_for(std::chrono::milliseconds(1000));
}
// Show cursor
show_console_cursor(true);
return 0;
}Indeterminate Progress Bar
You might have a use-case for a progress bar where the maximum amount of progress is unknown, e.g., you're downloading from a remote server that isn't advertising the total bytes.
Use an indicators::IndeterminateProgressBar for such cases. An IndeterminateProgressBar is similar to a regular progress bar except the total amount to progress towards is unknown. Ticking on this progress bar will happily run forever.
When you know progress is complete, simply call bar.mark_as_completed().
#include <chrono>
#include <indicators/indeterminate_progress_bar.hpp>
#include <indicators/cursor_control.hpp>
#include <indicators/termcolor.hpp>
#include <thread>
int main() {
indicators::IndeterminateProgressBar bar{
indicators::option::BarWidth{40},
indicators::option::Start{"["},
indicators::option::Fill{"·"},
indicators::option::Lead{"<==>"},
indicators::option::End{"]"},
indicators::option::PostfixText{"Checking for Updates"},
indicators::option::ForegroundColor{indicators::Color::yellow},
indicators::option::FontStyles{
std::vector<indicators::FontStyle>{indicators::FontStyle::bold}}
};
indicators::show_console_cursor(false);
auto job = [&bar]() {
std::this_thread::sleep_for(std::chrono::milliseconds(10000));
bar.mark_as_completed();
std::cout << termcolor::bold << termcolor::green
<< "System is up to date!\n" << termcolor::reset;
};
std::thread job_completion_thread(job);
// Update bar state
while (!bar.is_completed()) {
bar.tick();
std::this_thread::sleep_for(std::chrono::milliseconds(100));
}
job_completion_thread.join();
indicators::show_console_cursor(true);
return 0;
}Block Progress Bar
Are you in need of a smooth block progress bar using unicode block elements? Use BlockProgressBar instead of ProgressBar. Thanks to this blog post for making BlockProgressBar an easy addition to the library.
#include <indicators/block_progress_bar.hpp>
#include <thread>
#include <chrono>
int main() {
using namespace indicators;
// Hide cursor
show_console_cursor(false);
BlockProgressBar bar{
option::BarWidth{80},
option::Start{"["},
option::End{"]"},
option::ForegroundColor{Color::white} ,
option::FontStyles{std::vector<FontStyle>{FontStyle::bold}}
};
// Update bar state
auto progress = 0.0f;
while (true) {
bar.set_progress(progress);
progress += 0.25f;
if (bar.is_completed())
break;
std::this_thread::sleep_for(std::chrono::milliseconds(50));
}
// Show cursor
show_console_cursor(true);
return 0;
}MultiProgress
indicators supports management of multiple progress bars with the MultiProgress class template.
template <typename Indicator, size_t count> class MultiProgress is a class template that holds references to multiple progress bars and provides a safe interface to update the state of each bar. MultiProgress works with both ProgressBar and BlockProgressBar classes.
Use this class if you know the number of progress bars to manage at compile time.
Below is an example MultiProgress object that manages three ProgressBar objects.
#include <indicators/multi_progress.hpp>
#include <indicators/progress_bar.hpp>
int main() {
using namespace indicators;
// Configure first progress bar
ProgressBar bar1{
option::BarWidth{50},
option::Start{"["},
option::Fill{"■"},
option::Lead{"■"},
option::Remainder{" "},
option::End{" ]"},
option::ForegroundColor{Color::yellow},
option::ShowElapsedTime{true},
option::ShowRemainingTime{true},
option::PrefixText{"Progress Bar #1 "},
option::FontStyles{std::vector<FontStyle>{FontStyle::bold}}
};
// Configure second progress bar
ProgressBar bar2{
option::BarWidth{50},
option::Start{"["},
option::Fill{"="},
option::Lead{">"},
option::Remainder{" "},
option::End{" ]"},
option::ForegroundColor{Color::cyan},
option::ShowElapsedTime{true},
option::ShowRemainingTime{true},
option::PrefixText{"Progress Bar #2 "},
option::FontStyles{std::vector<FontStyle>{FontStyle::bold}}
};
// Configure third progress bar
indicators::ProgressBar bar3{
option::BarWidth{50},
option::Start{"["},
option::Fill{"#"},
option::Lead{"#"},
option::Remainder{" "},
option::End{" ]"},
option::ForegroundColor{Color::red},
option::ShowElapsedTime{true},
option::ShowRemainingTime{true},
option::PrefixText{"Progress Bar #3 "},
option::FontStyles{std::vector<FontStyle>{FontStyle::bold}}
};
// Construct MultiProgress object
indicators::MultiProgress<indicators::ProgressBar, 3> bars(bar1, bar2, bar3);
std::cout << "Multiple Progress Bars:\n";
auto job1 = [&bars]() {
while (true) {
bars.tick<0>();
if (bars.is_completed<0>())
break;
std::this_thread::sleep_for(std::chrono::milliseconds(100));
}
};
auto job2 = [&bars]() {
while (true) {
bars.tick<1>();
if (bars.is_completed<1>())
break;
std::this_thread::sleep_for(std::chrono::milliseconds(200));
}
};
auto job3 = [&bars]() {
while (true) {
bars.tick<2>();
if (bars.is_completed<2>())
break;
std::this_thread::sleep_for(std::chrono::milliseconds(60));
}
};
std::thread first_job(job1);
std::thread second_job(job2);
std::thread third_job(job3);
first_job.join();
second_job.join();
third_job.join();
return 0;
}DynamicProgress
DynamicProgress is a container class, similar to MultiProgress, for managing multiple progress bars. As the name suggests, with DynamicProgress, you can dynamically add new progress bars.
To add new progress bars, call bars.push_back(new_bar). This call will return the index of the appended bar. You can then refer to this bar with the indexing operator, e.g., bars[4].set_progress(55).
Use this class if you don't know the number of progress bars at compile time.
Below is an example DynamicProgress object that manages six ProgressBar objects. Three of these bars are added dynamically.
#include <indicators/dynamic_progress.hpp>
#include <indicators/progress_bar.hpp>
using namespace indicators;
int main() {
ProgressBar bar1{option::BarWidth{50}, option::ForegroundColor{Color::red},
option::ShowElapsedTime{true}, option::ShowRemainingTime{true},
option::PrefixText{"5c90d4a2d1a8: Downloading "}};
ProgressBar bar2{option::BarWidth{50}, option::ForegroundColor{Color::yellow},
option::ShowElapsedTime{true}, option::ShowRemainingTime{true},
option::PrefixText{"22337bfd13a9: Downloading "}};
ProgressBar bar3{option::BarWidth{50}, option::ForegroundColor{Color::green},
option::ShowElapsedTime{true}, option::ShowRemainingTime{true},
option::PrefixText{"10f26c680a34: Downloading "}};
ProgressBar bar4{option::BarWidth{50}, option::ForegroundColor{Color::white},
option::ShowElapsedTime{true}, option::ShowRemainingTime{true},
option::PrefixText{"6364e0d7a283: Downloading "}};
ProgressBar bar5{option::BarWidth{50}, option::ForegroundColor{Color::blue},
option::ShowElapsedTime{true}, option::ShowRemainingTime{true},
option::PrefixText{"ff1356ba118b: Downloading "}};
ProgressBar bar6{option::BarWidth{50}, option::ForegroundColor{Color::cyan},
option::ShowElapsedTime{true}, option::ShowRemainingTime{true},
option::PrefixText{"5a17453338b4: Downloading "}};
std::cout << termcolor::bold << termcolor::white << "Pulling image foo:bar/baz\n";
// Construct with 3 progress bars. We'll add 3 more at a later point
DynamicProgress<ProgressBar> bars(bar1, bar2, bar3);
// Do not hide bars when completed
bars.set_option(option::HideBarWhenComplete{false});
std::thread fourth_job, fifth_job, sixth_job;
auto job4 = [&bars](size_t i) {
while (true) {
bars[i].tick();
if (bars[i].is_completed()) {
bars[i].set_option(option::PrefixText{"6364e0d7a283: Pull complete "});
bars[i].mark_as_completed();
break;
}
std::this_thread::sleep_for(std::chrono::milliseconds(50));
}
};
auto job5 = [&bars](size_t i) {
while (true) {
bars[i].tick();
if (bars[i].is_completed()) {
bars[i].set_option(option::PrefixText{"ff1356ba118b: Pull complete "});
bars[i].mark_as_completed();
break;
}
std::this_thread::sleep_for(std::chrono::milliseconds(100));
}
};
auto job6 = [&bars](size_t i) {
while (true) {
bars[i].tick();
if (bars[i].is_completed()) {
bars[i].set_option(option::PrefixText{"5a17453338b4: Pull complete "});
bars[i].mark_as_completed();
break;
}
std::this_thread::sleep_for(std::chrono::milliseconds(40));
}
};
auto job1 = [&bars, &bar6, &sixth_job, &job6]() {
while (true) {
bars[0].tick();
if (bars[0].is_completed()) {
bars[0].set_option(option::PrefixText{"5c90d4a2d1a8: Pull complete "});
// bar1 is completed, adding bar6
auto i = bars.push_back(bar6);
sixth_job = std::thread(job6, i);
sixth_job.join();
break;
}
std::this_thread::sleep_for(std::chrono::milliseconds(140));
}
};
auto job2 = [&bars, &bar5, &fifth_job, &job5]() {
while (true) {
bars[1].tick();
if (bars[1].is_completed()) {
bars[1].set_option(option::PrefixText{"22337bfd13a9: Pull complete "});
// bar2 is completed, adding bar5
auto i = bars.push_back(bar5);
fifth_job = std::thread(job5, i);
fifth_job.join();
break;
}
std::this_thread::sleep_for(std::chrono::milliseconds(25));
}
};
auto job3 = [&bars, &bar4, &fourth_job, &job4]() {
while (true) {
bars[2].tick();
if (bars[2].is_completed()) {
bars[2].set_option(option::PrefixText{"10f26c680a34: Pull complete "});
// bar3 is completed, adding bar4
auto i = bars.push_back(bar4);
fourth_job = std::thread(job4, i);
fourth_job.join();
break;
}
std::this_thread::sleep_for(std::chrono::milliseconds(50));
}
};
std::thread first_job(job1);
std::thread second_job(job2);
std::thread third_job(job3);
third_job.join();
second_job.join();
first_job.join();
std::cout << termcolor::bold << termcolor::green << "✔ Downloaded image foo/bar:baz" << std::endl;
std::cout << termcolor::reset;
return 0;
}In the above code, notice the option bars.set_option(option::HideBarWhenComplete{true});. Yes, you can hide progress bars as and when they complete by setting this option to true. If you do so, the above example will look like this:
Progress Spinner
To introduce a progress spinner in your application, include indicators/progress_spinner.hpp and create a ProgressSpinner object. Here's the general structure of a progress spinner:
{prefix} {spinner} {percentage} [{elapsed}<{remaining}] {postfix}
ProgressSpinner has a vector of strings: spinner_states. At each update, the spinner will pick the next string from this sequence to print to the console. The spinner state can be updated similarly to ProgressBars: Using either tick() or set_progress(value).
#include <indicators/progress_spinner.hpp>
int main() {
using namespace indicators;
indicators::ProgressSpinner spinner{
option::PostfixText{"Checking credentials"},
option::ForegroundColor{Color::yellow},
option::SpinnerStates{std::vector<std::string>{"⠈", "⠐", "⠠", "⢀", "⡀", "⠄", "⠂", "⠁"}},
option::FontStyles{std::vector<FontStyle>{FontStyle::bold}}
};
// Update spinner state
auto job = [&spinner]() {
while (true) {
if (spinner.is_completed()) {
spinner.set_option(option::ForegroundColor{Color::green});
spinner.set_option(option::PrefixText{"✔"});
spinner.set_option(option::ShowSpinner{false});
spinner.set_option(option::ShowPercentage{false});
spinner.set_option(option::PostfixText{"Authenticated!"});
spinner.mark_as_completed();
break;
} else
spinner.tick();
std::this_thread::sleep_for(std::chrono::milliseconds(40));
}
};
std::thread thread(job);
thread.join();
return 0;
}Decremental Progress
indicators allows you to easily control the progress direction, i.e., incremental or decremental progress by using option::ProgressType. To program a countdown progress bar, use option::ProgressType::decremental
#include <chrono>
#include <indicators/progress_bar.hpp>
#include <thread>
using namespace indicators;
int main() {
ProgressBar bar{option::BarWidth{50},
option::ProgressType{ProgressType::decremental},
option::Start{"["},
option::Fill{"■"},
option::Lead{"■"},
option::Remainder{"-"},
option::End{"]"},
option::PostfixText{"Reverting System Restore"},
option::ForegroundColor{Color::yellow},
option::FontStyles{std::vector<FontStyle>{FontStyle::bold}}};
// Update bar state
while (true) {
bar.tick();
if (bar.is_completed())
break;
std::this_thread::sleep_for(std::chrono::milliseconds(100));
}
std::cout << termcolor::bold << termcolor::white
<< "Task Failed Successfully\n" << termcolor::reset;
return 0;
}Working with Iterables
If you'd like to use progress bars to indicate progress while iterating over iterables, e.g., a list of numbers, this can be achieved by using the option::MaxProgress:
#include <chrono>
#include <indicators/block_progress_bar.hpp>
#include <indicators/cursor_control.hpp>
#include <thread>
int main() {
// Hide cursor
indicators::show_console_cursor(false);
// Random list of numbers
std::vector<size_t> numbers;
for (size_t i = 0; i < 1259438; ++i) {
numbers.push_back(i);
}
using namespace indicators;
BlockProgressBar bar{
option::BarWidth{80},
option::ForegroundColor{Color::white},
option::FontStyles{
std::vector<FontStyle>{FontStyle::bold}},
option::MaxProgress{numbers.size()}
};
std::cout << "Iterating over a list of numbers (size = "
<< numbers.size() << ")\n";
std::vector<size_t> result;
for (size_t i = 0; i < numbers.size(); ++i) {
// Perform some computation
result.push_back(numbers[i] * numbers[i]);
// Show iteration as postfix text
bar.set_option(option::PostfixText{
std::to_string(i) + "/" + std::to_string(numbers.size())
});
// update progress bar
bar.tick();
}
bar.mark_as_completed();
// Show cursor
indicators::show_console_cursor(true);
return 0;
}Unicode Support
indicators supports multi-byte unicode characters in progress bars.
If the option::BarWidth is set, the library aims to respect this setting. When filling the bar, if the next Fill string has a display width that would exceed the bar width, then the library will fill the remainder of the bar with ' ' space characters instead.
See below an example of some progress bars, each with a bar width of 50, displaying different unicode characters:
#include <chrono>
#include <indicators/progress_bar.hpp>
#include <indicators/indeterminate_progress_bar.hpp>
#include <indicators/cursor_control.hpp>
#include <thread>
int main() {
indicators::show_console_cursor(false);
std::this_thread::sleep_for(std::chrono::milliseconds(2000));
{
// Plain old ASCII
indicators::ProgressBar bar{
indicators::option::BarWidth{50},
indicators::option::Start{"["},
indicators::option::Fill{"="},
indicators::option::Lead{">"},
indicators::option::Remainder{" "},
indicators::option::End{" ]"},
indicators::option::PostfixText{"Plain-old ASCII"},
indicators::option::ForegroundColor{indicators::Color::green},
indicators::option::FontStyles{
std::vector<indicators::FontStyle>{indicators::FontStyle::bold}}
};
// Update bar state
while (true) {
bar.tick();
if (bar.is_completed())
break;
std::this_thread::sleep_for(std::chrono::milliseconds(10));
}
}
{
// Unicode
indicators::ProgressBar bar{
indicators::option::BarWidth{50},
indicators::option::Start{"["},
indicators::option::Fill{"驚くばかり"},
indicators::option::Lead{">"},
indicators::option::Remainder{" "},
indicators::option::End{" ]"},
indicators::option::PostfixText{"Japanese"},
indicators::option::ForegroundColor{indicators::Color::yellow},
indicators::option::FontStyles{
std::vector<indicators::FontStyle>{indicators::FontStyle::bold}}
};
// Update bar state
while (true) {
bar.tick();
if (bar.is_completed())
break;
std::this_thread::sleep_for(std::chrono::milliseconds(10));
}
}
{
// Russian
indicators::ProgressBar bar{
indicators::option::BarWidth{50},
indicators::option::Start{"["},
indicators::option::Fill{"Потрясающие"},
indicators::option::Remainder{" "},
indicators::option::End{" ]"},
indicators::option::PostfixText{"Russian"},
indicators::option::ForegroundColor{indicators::Color::red},
indicators::option::FontStyles{
std::vector<indicators::FontStyle>{indicators::FontStyle::bold}}
};
// Update bar state
while (true) {
bar.tick();
if (bar.is_completed())
break;
std::this_thread::sleep_for(std::chrono::milliseconds(10));
}
}
{
// Greek
indicators::ProgressBar bar{
indicators::option::BarWidth{50},
indicators::option::Start{"["},
indicators::option::Fill{"Φοβερός"},
indicators::option::Remainder{" "},
indicators::option::End{" ]"},
indicators::option::PostfixText{"Greek"},
indicators::option::ForegroundColor{indicators::Color::cyan},
indicators::option::FontStyles{
std::vector<indicators::FontStyle>{indicators::FontStyle::bold}}
};
// Update bar state
while (true) {
bar.tick();
if (bar.is_completed())
break;
std::this_thread::sleep_for(std::chrono::milliseconds(10));
}
}
{
// Chinese
indicators::ProgressBar bar{
indicators::option::BarWidth{50},
indicators::option::Start{"["},
indicators::option::Fill{"太棒了"},
indicators::option::Remainder{" "},
indicators::option::End{" ]"},
indicators::option::PostfixText{"Chinese"},
indicators::option::ForegroundColor{indicators::Color::green},
indicators::option::FontStyles{
std::vector<indicators::FontStyle>{indicators::FontStyle::bold}}
};
// Update bar state
while (true) {
bar.tick();
if (bar.is_completed())
break;
std::this_thread::sleep_for(std::chrono::milliseconds(10));
}
}
{
// Emojis
indicators::ProgressBar bar{
indicators::option::BarWidth{50},
indicators::option::Start{"["},
indicators::option::Fill{"🔥"},
indicators::option::Lead{"🔥"},
indicators::option::Remainder{" "},
indicators::option::End{" ]"},
indicators::option::PostfixText{"Emojis"},
indicators::option::ForegroundColor{indicators::Color::white},
indicators::option::FontStyles{
std::vector<indicators::FontStyle>{indicators::FontStyle::bold}}
};
// Update bar state
while (true) {
bar.tick();
if (bar.is_completed())
break;
std::this_thread::sleep_for(std::chrono::milliseconds(10));
}
}
{
// Indeterminate progress bar
indicators::IndeterminateProgressBar bar{
indicators::option::BarWidth{50},
indicators::option::Start{"["},
indicators::option::Fill{"✯"},
indicators::option::Lead{"載入中"},
indicators::option::End{" ]"},
indicators::option::PostfixText{"Loading Progress Bar"},
indicators::option::ForegroundColor{indicators::Color::yellow},
indicators::option::FontStyles{
std::vector<indicators::FontStyle>{indicators::FontStyle::bold}}
};
auto job = [&bar]() {
std::this_thread::sleep_for(std::chrono::milliseconds(10000));
bar.mark_as_completed();
};
std::thread job_completion_thread(job);
// Update bar state
while (!bar.is_completed()) {
bar.tick();
std::this_thread::sleep_for(std::chrono::milliseconds(100));
}
job_completion_thread.join();
}
indicators::show_console_cursor(true);
return 0;
}Building Samples
git clone https://github.com/p-ranav/indicators
cd indicators
mkdir build && cd build
cmake -DINDICATORS_SAMPLES=ON -DINDICATORS_DEMO=ON ..
makeWinLibs + MinGW
For Windows, if you use WinLibs like I do, the cmake command would look like this:
foo@bar:~$ mkdir build && cd build
foo@bar:~$ cmake -G "MinGW Makefiles" -DCMAKE_CXX_COMPILER="C:/WinLibs/mingw64/bin/g++.exe" -DINDICATORS_SAMPLES=ON -DINDICATORS_DEMO=ON ..
foo@bar:~$ make -j4Generating Single Header
python3 utils/amalgamate/amalgamate.py -c single_include.json -s .Contributing
Contributions are welcome, have a look at the CONTRIBUTING.md document for more information.
Author: p-ranav
Source Code: https://github.com/p-ranav/indicators/
License: View license
1651161600
Curlcpp: An Object Oriented C++ Wrapper for CURL (libcurl)
curlcpp
An object-oriented C++ wrapper for cURL tool
If you want to know a bit more about cURL and libcurl, you should go on the official website http://curl.haxx.se/
Donate
Help me to improve this project!
Compile and link
Standalone
cd build
cmake ..
make # -j2
Note: cURL >= 7.34 is required.
Then add <curlcpp root>/build/src/ to your library path and <curlcpp root>/include/ to your include path.
When linking, link against curlcpp (e.g.: g++ -std=c++11 example.cpp -o example -lcurlcpp -lcurl). Or if you want run from terminal,
g++ -std=c++11 example.cpp -L/home/username/path/to/build/src/ -I/home/username/path/to/include/ -lcurlcpp -lcurl
Submodule
When using a git submodule and CMake-buildsystem, add the following lines to your CMakeLists.txt:
ADD_SUBDIRECTORY(ext/curlcpp) # Change `ext/curlcpp` to a directory according to your setup
INCLUDE_DIRECTORIES(${CURLCPP_SOURCE_DIR}/include)
Simple usage example
Here are some usage examples. You will find more examples in the test folder!
Here's an example of a simple HTTP request to get google web page, using the curl_easy interface:
#include "curl_easy.h"
using curl::curl_easy;
using curl::curl_easy_exception;
using curl::curlcpp_traceback;
/**
* This example shows how to make a simple request with curl.
*/
int main() {
// Easy object to handle the connection.
curl_easy easy;
// Add some options.
easy.add<CURLOPT_URL>("http://<your_url_here>");
easy.add<CURLOPT_FOLLOWLOCATION>(1L);
try {
easy.perform();
} catch (curl_easy_exception &error) {
// If you want to print the last error.
std::cerr<<error.what()<<std::endl;
}
return 0;
}
If you want to get information about the current curl session, you could do:
#include "curl_easy.h"
#include "curl_ios.h"
#include "curl_exception.h"
using std::ostringstream;
using curl::curl_easy;
using curl::curl_easy_exception;
using curl::curlcpp_traceback;
using curl::curl_ios;
/**
* This example shows how to use the easy interface and obtain
* informations about the current session.
*/
int main(int argc, const char **argv) {
// Let's declare a stream
ostringstream stream;
// We are going to put the request's output in the previously declared stream
curl_ios<ostringstream> ios(stream);
// Declaration of an easy object
curl_easy easy(ios);
// Add some option to the curl_easy object.
easy.add<CURLOPT_URL>("http://<your_url_here>");
easy.add<CURLOPT_FOLLOWLOCATION>(1L);
try {
easy.perform();
// Retrieve information about curl current session.
auto x = easy.get_info<CURLINFO_CONTENT_TYPE>();
/**
* get_info returns a curl_easy_info object. With the get method we retrieve
* the std::pair object associated with it: the first item is the return code of the
* request. The second is the element requested by the specified libcurl macro.
*/
std::cout<<x.get()<<std::endl;
} catch (curl_easy_exception &error) {
// If you want to print the last error.
std::cerr<<error.what()<<std::endl;
// If you want to print the entire error stack you can do
error.print_traceback();
}
return 0;
}
Here's instead, the creation of an HTTPS POST login form:
#include <string>
#include "curl_easy.h"
#include "curl_pair.h"
#include "curl_form.h"
#include "curl_exception.h"
using std::string;
using curl::curl_form;
using curl::curl_easy;
using curl::curl_pair;
using curl::curl_easy_exception;
using curl::curlcpp_traceback;
int main(int argc, const char * argv[]) {
curl_form form;
curl_easy easy;
// Forms creation
curl_pair<CURLformoption,string> name_form(CURLFORM_COPYNAME,"user");
curl_pair<CURLformoption,string> name_cont(CURLFORM_COPYCONTENTS,"you username here");
curl_pair<CURLformoption,string> pass_form(CURLFORM_COPYNAME,"passw");
curl_pair<CURLformoption,string> pass_cont(CURLFORM_COPYCONTENTS,"your password here");
try {
// Form adding
form.add(name_form,name_cont);
form.add(pass_form,pass_cont);
// Add some options to our request
easy.add<CURLOPT_URL>("http://<your_url_here>");
easy.add<CURLOPT_SSL_VERIFYPEER>(false);
easy.add<CURLOPT_HTTPPOST>(form.get());
// Execute the request.
easy.perform();
} catch (curl_easy_exception &error) {
// If you want to get the entire error stack we can do:
curlcpp_traceback errors = error.get_traceback();
// Otherwise we could print the stack like this:
error.print_traceback();
}
return 0;
}
And if we would like to put the returned content in a file? Nothing easier than:
#include <iostream>
#include <ostream>
#include <fstream>
#include "curl_easy.h"
#include "curl_ios.h"
#include "curl_exception.h"
using std::cout;
using std::endl;
using std::ostream;
using std::ofstream;
using curl::curl_easy;
using curl::curl_ios;
using curl::curl_easy_exception;
using curl::curlcpp_traceback;
int main(int argc, const char * argv[]) {
// Create a file
ofstream myfile;
myfile.open ("/path/to/your/file");
// Create a curl_ios object to handle the stream
curl_ios<ostream> writer(myfile);
// Pass it to the easy constructor and watch the content returned in that file!
curl_easy easy(writer);
// Add some option to the easy handle
easy.add<CURLOPT_URL>("http://<your_url_here>");
easy.add<CURLOPT_FOLLOWLOCATION>(1L);
try {
// Execute the request
easy.perform();
} catch (curl_easy_exception &error) {
// If you want to print the last error.
std::cerr<<error.what()<<std::endl;
// If you want to print the entire error stack you can do
error.print_traceback();
}
myfile.close();
return 0;
}
Not interested in files? So let's put the request's output in a variable!
#include <iostream>
#include <ostream>
#include "curl_easy.h"
#include "curl_form.h"
#include "curl_ios.h"
#include "curl_exception.h"
using std::cout;
using std::endl;
using std::ostringstream;
using curl::curl_easy;
using curl::curl_ios;
using curl::curl_easy_exception;
using curl::curlcpp_traceback;
int main() {
// Create a stringstream object
ostringstream str;
// Create a curl_ios object, passing the stream object.
curl_ios<ostringstream> writer(str);
// Pass the writer to the easy constructor and watch the content returned in that variable!
curl_easy easy(writer);
// Add some option to the easy handle
easy.add<CURLOPT_URL>("http://<your_url_here>");
easy.add<CURLOPT_FOLLOWLOCATION>(1L);
try {
easy.perform();
// Let's print the stream content
cout<<str.str()<<endl;
} catch (curl_easy_exception &error) {
// If you want to print the last error.
std::cerr<<error.what()<<std::endl;
// If you want to print the entire error stack you can do
error.print_traceback();
}
return 0;
}
I have implemented a sender and a receiver to make it easy to use send/receive without handling buffers. For example, a very simple send/receiver would be:
#include <iostream>
#include <string>
#include "curl_easy.h"
#include "curl_form.h"
#include "curl_pair.h"
#include "curl_receiver.h"
#include "curl_exception.h"
#include "curl_sender.h"
using std::cout;
using std::endl;
using std::string;
using curl::curl_form;
using curl::curl_easy;
using curl::curl_sender;
using curl::curl_receiver;
using curl::curl_easy_exception;
using curl::curlcpp_traceback;
int main(int argc, const char * argv[]) {
// Simple request
string request = "GET / HTTP/1.0\r\nHost: example.com\r\n\r\n";
// Creation of easy object.
curl_easy easy;
try {
easy.add<CURLOPT_URL>("http://<your_url_here>");
// Just connect
easy.add<CURLOPT_CONNECT_ONLY>(true);
// Execute the request.
easy.perform();
} catch (curl_easy_exception &error) {
// If you want to get the entire error stack we can do:
curlcpp_traceback errors = error.get_traceback();
// Otherwise we could print the stack like this:
error.print_traceback();
}
// Creation of a sender. You should wait here using select to check if socket is ready to send.
curl_sender<string> sender(easy);
sender.send(request);
// Prints che sent bytes number.
cout<<"Sent bytes: "<<sender.get_sent_bytes()<<endl;
for(;;) {
// You should wait here to check if socket is ready to receive
try {
// Create a receiver
curl_receiver<char, 1024> receiver;
// Receive the content on the easy handler
receiver.receive(easy);
// Prints the received bytes number.
cout<<"Receiver bytes: "<<receiver.get_received_bytes()<<endl;
} catch (curl_easy_exception &error) {
// If any errors occurs, exit from the loop
break;
}
}
return 0;
}Author: JosephP91
Source Code: https://github.com/JosephP91/curlcpp
License: MIT License
#cpluplus #c
1649659620
Concurrencpp: Modern Concurrency for C++.
concurrencpp, the C++ concurrency library
concurrencpp is a tasking library for C++ allowing developers to write highly concurrent applications easily and safely by using tasks, executors and coroutines. By using concurrencpp applications can break down big procedures that need to be processed asynchronously into smaller tasks that run concurrently and work in a co-operative manner to achieve the wanted result. concurrencpp also allows applications to write parallel algorithms easily by using parallel coroutines.
concurrencpp main advantages are:
- Being able to write modern concurrency code without having to rely on low-level concurrency primitives like locks and condition variables.
- Being able to write highly concurrent and parallel applications that scale automatically to use all hardware resources, as needed.
- Being able to write non-blocking, synchronous-like code easily by using C++20 coroutines and the
co_awaitkeyword. - Reducing the possibility of race conditions, data races and deadlocks by using high-level objects with built-in synchronization.
- concurrencpp provides various types of commonly used executors with a complete coroutine integration.
- Applications can extend the library by implementing their own provided executors.
Table of contents
- concurrencpp overview
- Tasks
- Executors
- Result objects
- Parallel coroutines
- Result-promises
- Shared result objects
- Termination in concurrencpp
- Resume executors
- Utility functions
- Timers and Timer queues
- Generators
- The runtime object
- Supported platforms and tools
- Building, installing and testing
concurrencpp overview
concurrencpp is a task-centric library. A task is an asynchronous operation. Tasks offer a higher level of abstraction for concurrent code than traditional thread-centric approaches. Tasks can be chained together, meaning that tasks pass their asynchronous result from one to another, where the result of one task is used as if it were a parameter or an intermediate value of another ongoing task. Tasks allow applications to utilize available hardware resources better and scale much more than using raw threads, since tasks can be suspended, waiting for another task to produce a result, without blocking underlying OS-threads. Tasks bring much more productivity to developers by allowing them to focus more on business-logic and less on low-level concepts like thread management and inter-thread synchronization.
While tasks specify what actions have to be executed, executors are worker-objects that specify where and how to execute tasks. Executors spare applications the managing of thread pools and task queues themselves. Executors also decouple those concepts away from application code, by providing a unified API for creating and scheduling tasks.
Tasks communicate with each other using result objects. A result object is an asynchronous pipe that pass the asynchronous result of one task to another ongoing-task. Results can be awaited and resolved in a non-blocking manner.
These three concepts - the task, the executor and the associated result are the building blocks of concurrencpp. Executors run tasks that communicate with each-other by sending results through result-objects. Tasks, executors and result objects work together symbiotically to produce concurrent code which is fast and clean.
concurrencpp is built around the RAII concept. In order to use tasks and executors, applications create a runtime instance in the beginning of the main function. The runtime is then used to acquire existing executors and register new user-defined executors. Executors are used to create and schedule tasks to run, and they might return a result object that can be used to marshal the asynchronous result to another task that acts as its consumer. When the runtime is destroyed, it iterates over every stored executor and calls its shutdown method. Every executor then exits gracefully. Unscheduled tasks are destroyed, and attempts to create new tasks will throw an exception.
"Hello world" program using concurrencpp:
#include "concurrencpp/concurrencpp.h"
#include <iostream>
int main() {
concurrencpp::runtime runtime;
auto result = runtime.thread_executor()->submit([] {
std::cout << "hello world" << std::endl;
});
result.get();
return 0;
}
In this basic example, we created a runtime object, then we acquired the thread executor from the runtime. We used submit to pass a lambda as our given callable. This lambda returns void, hence, the executor returns a result<void> object that marshals the asynchronous result back to the caller. main calls get which blocks the main thread until the result becomes ready. If no exception was thrown, get returns void. If an exception was thrown, get re-throws it. Asynchronously, thread_executor launches a new thread of execution and runs the given lambda. It implicitly co_return void and the task is finished. main is then unblocked.
Concurrent even-number counting:
#include "concurrencpp/concurrencpp.h"
#include <iostream>
#include <vector>
#include <algorithm>
#include <ctime>
using namespace concurrencpp;
std::vector<int> make_random_vector() {
std::vector<int> vec(64 * 1'024);
std::srand(std::time(nullptr));
for (auto& i : vec) {
i = ::rand();
}
return vec;
}
result<size_t> count_even(std::shared_ptr<thread_pool_executor> tpe, const std::vector<int>& vector) {
const auto vecor_size = vector.size();
const auto concurrency_level = tpe->max_concurrency_level();
const auto chunk_size = vecor_size / concurrency_level;
std::vector<result<size_t>> chunk_count;
for (auto i = 0; i < concurrency_level; i++) {
const auto chunk_begin = i * chunk_size;
const auto chunk_end = chunk_begin + chunk_size;
auto result = tpe->submit([&vector, chunk_begin, chunk_end]() -> size_t {
return std::count_if(vector.begin() + chunk_begin, vector.begin() + chunk_end, [](auto i) {
return i % 2 == 0;
});
});
chunk_count.emplace_back(std::move(result));
}
size_t total_count = 0;
for (auto& result : chunk_count) {
total_count += co_await result;
}
co_return total_count;
}
int main() {
concurrencpp::runtime runtime;
const auto vector = make_random_vector();
auto result = count_even(runtime.thread_pool_executor(), vector);
const auto total_count = result.get();
std::cout << "there are " << total_count << " even numbers in the vector" << std::endl;
return 0;
}
In this example, we start the program by creating a runtime object. We create a vector filled with random numbers, then we acquire the thread_pool_executor from the runtime and call count_even. count_even is a coroutine that spawns more tasks and co_awaits for them to finish inside. max_concurrency_level returns the maximum amount of workers that the executor supports, In the threadpool executor case, the number of workers is calculated from the number of cores. We then partition the array to match the number of workers and send every chunk to be processed in its own task. Asynchronously, the workers count how many even numbers each chunk contains, and co_return the result. count_even sums every result by pulling the count using co_await, the final result is then co_returned. The main thread, which was blocked by calling get is unblocked and the total count is returned. main prints the number of even numbers and the program terminates gracefully.
Tasks
Every big or complex operation can be broken down to smaller and chainable steps. Tasks are asynchronous operations implementing those computational steps. Tasks can run anywhere with the help of executors. While tasks can be created from regular callables (such as functors and lambdas), Tasks are mostly used with coroutines, which allow smooth suspension and resumption. In concurrencpp, the task concept is represented by the concurrencpp::task class. Although the task concept is central to concurrenpp, applications will rarely have to create and manipulate task objects themselves, as task objects are created and scheduled by the runtime with no external help.
concurrencpp coroutines
concurrencpp allows applications to produce and consume coroutines as the main way of creating tasks. concurrencpp supports both eager and lazy tasks.
Eager tasks start to run the moment they are invoked. This type of execution is recommended when applications need to fire an asynchronous action and consume its result later on (fire and consume later), or completely ignore the asynchronous result (fire and forget).
Eager tasks can return result or null_result. result return type tells the coroutine to marshal the returned value or the thrown exception (fire and consume later) while null_result return type tells the coroutine to drop and ignore any of them (fire and forget).
Eager coroutines can start to run synchronously, in the caller thread. This kind of coroutines is called "regular coroutines". Concurrencpp eager coroutines can also start to run in parallel, inside a given executor, this kind of coroutines is called "parallel coroutines".
Lazy tasks, on the other hand, start to run only when co_awaited. This type of tasks is recommended when the result of the task is meant to be consumed immediately after creating the task. Lazy tasks, being deferred, are a bit more optimized for the case of immediate-consumption, as they do not need special thread-synchronization in order to marshal the asynchronous result back to its consumer. The compiler might also optimize away some memory allocations needed to form the underlying coroutine promise. It is not possible to fire a lazy task and execute something else meanwhile - the firing of a lazy-callee coroutine necessarily means the suspension of the caller-coroutine. The caller coroutine will only be resumed when the lazy-callee coroutine completes. Lazy tasks can only return lazy_result.
Lazy tasks can be converted to eager tasks by calling lazy_result::run. This method runs the lazy task inline and returns a result object that monitors the newly started task. If developers are unsure which result type to use, they are encouraged to use lazy results, as they can be converted to regular (eager) results if needed.
When a function returns any of lazy_result, result or null_resultand contains at least one co_await or co_return in its body, the function is a concurrencpp coroutine. Every valid concurrencpp coroutine is a valid task. In our count-even example above, count_even is such a coroutine. We first spawned count_even, then inside it the threadpool executor spawned more child tasks (that are created from regular callables), that were eventually joined using co_await.
Executors
A concurrencpp executor is an object that is able to schedule and run tasks. Executors simplify the work of managing resources such as threads, thread pools and task queues by decoupling them away from application code. Executors provide a unified way of scheduling and executing tasks, since they all extend concurrencpp::executor.
executor API
class executor {
/*
Initializes a new executor and gives it a name.
*/
executor(std::string_view name);
/*
Destroys this executor.
*/
virtual ~executor() noexcept = default;
/*
The name of the executor, used for logging and debugging.
*/
const std::string name;
/*
Schedules a task to run in this executor.
Throws concurrencpp::errors::runtime_shutdown exception if shutdown was called before.
*/
virtual void enqueue(concurrencpp::task task) = 0;
/*
Schedules a range of tasks to run in this executor.
Throws concurrencpp::errors::runtime_shutdown exception if shutdown was called before.
*/
virtual void enqueue(std::span<concurrencpp::task> tasks) = 0;
/*
Returns the maximum count of real OS threads this executor supports.
The actual count of threads this executor is running might be smaller than this number.
returns numeric_limits<int>::max if the executor does not have a limit for OS threads.
*/
virtual int max_concurrency_level() const noexcept = 0;
/*
Returns true if shutdown was called before, false otherwise.
*/
virtual bool shutdown_requested() const noexcept = 0;
/*
Shuts down the executor:
- Tells underlying threads to exit their work loop and joins them.
- Destroys unexecuted coroutines.
- Makes subsequent calls to enqueue, post, submit, bulk_post and
bulk_submit to throw concurrencpp::errors::runtime_shutdown exception.
- Makes shutdown_requested return true.
*/
virtual void shutdown() noexcept = 0;
/*
Turns a callable and its arguments into a task object and
schedules it to run in this executor using enqueue.
Arguments are passed to the task by decaying them first.
Throws errors::runtime_shutdown exception if shutdown has been called before.
*/
template<class callable_type, class ... argument_types>
void post(callable_type&& callable, argument_types&& ... arguments);
/*
Like post, but returns a result object that marshals the asynchronous result.
Throws errors::runtime_shutdown exception if shutdown has been called before.
*/
template<class callable_type, class ... argument_types>
result<type> submit(callable_type&& callable, argument_types&& ... arguments);
/*
Turns an array of callables into an array of tasks and
schedules them to run in this executor using enqueue.
Throws errors::runtime_shutdown exception if shutdown has been called before.
*/
template<class callable_type>
void bulk_post(std::span<callable_type> callable_list);
/*
Like bulk_post, but returns an array of result objects that marshal the asynchronous results.
Throws errors::runtime_shutdown exception if shutdown has been called before.
*/
template<class callable_type>
std::vector<concurrencpp::result<type>> bulk_submit(std::span<callable_type> callable_list);
};
Executor types
As mentioned above, concurrencpp provides commonly used executors. These executor types are:
thread pool executor - a general purpose executor that maintains a pool of threads. The thread pool executor is suitable for short cpu-bound tasks that don't block. Applications are encouraged to use this executor as the default executor for non-blocking tasks. The concurrencpp thread pool provides dynamic thread injection and dynamic work balancing.
background executor - a threadpool executor with a larger pool of threads. Suitable for launching short blocking tasks like file io and db queries. Important note: when consuming results this executor returned by calling submit and bulk_submit, it is important to switch execution using resume_on to a cpu-bound executor, in order to prevent cpu-bound tasks to be processed inside background_executor.
example:
auto result = background_executor.submit([] { /* some blocking action */ });
auto done_result = co_await result.resolve();
co_await resume_on(some_cpu_executor);
auto val = co_await done_result; // runs inside some_cpu_executor
thread executor - an executor that launches each enqueued task to run on a new thread of execution. Threads are not reused. This executor is good for long running tasks, like objects that run a work loop, or long blocking operations.
worker thread executor - a single thread executor that maintains a single task queue. Suitable when applications want a dedicated thread that executes many related tasks.
manual executor - an executor that does not execute coroutines by itself. Application code can execute previously enqueued tasks by manually invoking its execution methods.
derivable executor - a base class for user defined executors. Although inheriting directly from concurrencpp::executor is possible, derivable_executor uses the CRTP pattern that provides some optimization opportunities for the compiler.
inline executor - mainly used to override the behavior of other executors. Enqueuing a task is equivalent to invoking it inline.
Using executors
The bare mechanism of an executor is encapsulated in its enqueue method. This method enqueues a task for execution and has two overloads: One overload receives a single task object as an argument, and another that receives a span of task objects. The second overload is used to enqueue a batch of tasks. This allows better scheduling heuristics and decreased contention.
Applications don't have to rely on enqueue alone, concurrencpp::executor provides an API for scheduling user callables by converting them to task objects behind the scenes. Applications can request executors to return a result object that marshals the asynchronous result of the provided callable. This is done by calling executor::submit and executor::bulk_submit. submit gets a callable, and returns a result object. executor::bulk_submit gets a span of callables and returns a vectorof result objects in a similar way submit works. In many cases, applications are not interested in the asynchronous value or exception. In this case, applications can use executor:::post and executor::bulk_post to schedule a callable or a span of callables to be executed, but also tells the task to drop any returned value or thrown exception. Not marshaling the asynchronous result is faster than marshaling, but then we have no way of knowing the status or the result of the ongoing task.
post, bulk_post, submit and bulk_submit use enqueue behind the scenes for the underlying scheduling mechanism.
thread_pool_executor API
Aside from post, submit, bulk_post and bulk_submit, the thread_pool_executor provides these additional methods.
class thread_pool_executor {
/*
Returns the number of milliseconds each thread-pool worker
remains idle (lacks any task to execute) before exiting.
This constant can be set by passing a runtime_options object
to the constructor of the runtime class.
*/
std::chrono::milliseconds max_worker_idle_time() const noexcept;
};
manual_executor API
Aside from post, submit, bulk_post and bulk_submit, the manual_executor provides these additional methods.
class manual_executor {
/*
Destructor. Equivalent to clear.
*/
~manual_executor() noexcept;
/*
Returns the number of enqueued tasks at the moment of invocation.
This number can change quickly by the time the application handles it, it should be used as a hint.
This method is thread safe.
Might throw std::system_error if one of the underlying synchronization primitives throws.
*/
size_t size() const noexcept;
/*
Queries whether the executor is empty from tasks at the moment of invocation.
This value can change quickly by the time the application handles it, it should be used as a hint.
This method is thread safe.
Might throw std::system_error if one of the underlying synchronization primitives throws.
*/
bool empty() const noexcept;
/*
Clears the executor from any enqueued but yet to-be-executed tasks,
and returns the number of cleared tasks.
Tasks enqueued to this executor by (post_)submit method are resumed
and errors::broken_task exception is thrown inside them.
Ongoing tasks that are being executed by loop_once(_XXX) or loop(_XXX) are uneffected.
This method is thread safe.
Might throw std::system_error if one of the underlying synchronization primitives throws.
Throws errors::shutdown_exception if shutdown was called before.
*/
size_t clear();
/*
Tries to execute a single task. If at the moment of invocation the executor
is empty, the method does nothing.
Returns true if a task was executed, false otherwise.
This method is thread safe.
Might throw std::system_error if one of the underlying synchronization primitives throws.
Throws errors::shutdown_exception if shutdown was called before.
*/
bool loop_once();
/*
Tries to execute a single task.
This method returns when either a task was executed or max_waiting_time
(in milliseconds) has reached.
If max_waiting_time is 0, the method is equivalent to loop_once.
If shutdown is called from another thread, this method returns
and throws errors::shutdown_exception.
This method is thread safe.
Might throw std::system_error if one of the underlying synchronization primitives throws.
Throws errors::shutdown_exception if shutdown was called before.
*/
bool loop_once_for(std::chrono::milliseconds max_waiting_time);
/*
Tries to execute a single task.
This method returns when either a task was executed or timeout_time has reached.
If timeout_time has already expired, this method is equivalent to loop_once.
If shutdown is called from another thread, this method
returns and throws errors::shutdown_exception.
This method is thread safe.
Might throw std::system_error if one of the underlying synchronization primitives throws.
Throws errors::shutdown_exception if shutdown was called before.
*/
template<class clock_type, class duration_type>
bool loop_once_until(std::chrono::time_point<clock_type, duration_type> timeout_time);
/*
Tries to execute max_count enqueued tasks and returns the number of tasks that were executed.
This method does not wait: it returns when the executor
becomes empty from tasks or max_count tasks have been executed.
This method is thread safe.
Might throw std::system_error if one of the underlying synchronization primitives throws.
Throws errors::shutdown_exception if shutdown was called before.
*/
size_t loop(size_t max_count);
/*
Tries to execute max_count tasks.
This method returns when either max_count tasks were executed or a
total amount of max_waiting_time has passed.
If max_waiting_time is 0, the method is equivalent to loop.
Returns the actual amount of tasks that were executed.
If shutdown is called from another thread, this method returns
and throws errors::shutdown_exception.
This method is thread safe.
Might throw std::system_error if one of the underlying synchronization primitives throws.
Throws errors::shutdown_exception if shutdown was called before.
*/
size_t loop_for(size_t max_count, std::chrono::milliseconds max_waiting_time);
/*
Tries to execute max_count tasks.
This method returns when either max_count tasks were executed or timeout_time has reached.
If timeout_time has already expired, the method is equivalent to loop.
Returns the actual amount of tasks that were executed.
If shutdown is called from another thread, this method returns
and throws errors::shutdown_exception.
This method is thread safe.
Might throw std::system_error if one of the underlying synchronization primitives throws.
Throws errors::shutdown_exception if shutdown was called before.
*/
template<class clock_type, class duration_type>
size_t loop_until(size_t max_count, std::chrono::time_point<clock_type, duration_type> timeout_time);
/*
Waits for at least one task to be available for execution.
This method should be used as a hint,
as other threads (calling loop, for example) might empty the executor,
before this thread has a chance to do something with the newly enqueued tasks.
If shutdown is called from another thread, this method returns
and throws errors::shutdown_exception.
This method is thread safe.
Might throw std::system_error if one of the underlying synchronization primitives throws.
Throws errors::shutdown_exception if shutdown was called before.
*/
void wait_for_task();
/*
This method returns when one or more tasks are available for
execution or max_waiting_time has passed.
Returns true if at at least one task is available for execution, false otherwise.
This method should be used as a hint, as other threads (calling loop, for example)
might empty the executor, before this thread has a chance to do something
with the newly enqueued tasks.
If shutdown is called from another thread, this method
returns and throws errors::shutdown_exception.
This method is thread safe.
Might throw std::system_error if one of the underlying synchronization primitives throws.
Throws errors::shutdown_exception if shutdown was called before.
*/
bool wait_for_task_for(std::chrono::milliseconds max_waiting_time);
/*
This method returns when one or more tasks are available for execution or timeout_time has reached.
Returns true if at at least one task is available for execution, false otherwise.
This method should be used as a hint,
as other threads (calling loop, for example) might empty the executor,
before this thread has a chance to do something with the newly enqueued tasks.
If shutdown is called from another thread, this method
returns and throws errors::shutdown_exception.
This method is thread safe.
Might throw std::system_error if one of the underlying synchronization primitives throws.
Throws errors::shutdown_exception if shutdown was called before.
*/
template<class clock_type, class duration_type>
bool wait_for_task_until(std::chrono::time_point<clock_type, duration_type> timeout_time);
/*
This method returns when max_count or more tasks are available for execution.
This method should be used as a hint, as other threads
(calling loop, for example) might empty the executor,
before this thread has a chance to do something with the newly enqueued tasks.
If shutdown is called from another thread, this method returns
and throws errors::shutdown_exception.
This method is thread safe.
Might throw std::system_error if one of the underlying synchronization primitives throws.
Throws errors::shutdown_exception if shutdown was called before.
*/
void wait_for_tasks(size_t max_count);
/*
This method returns when max_count or more tasks are available for execution
or max_waiting_time (in milliseconds) has passed.
Returns the number of tasks available for execution when the method returns.
This method should be used as a hint, as other
threads (calling loop, for example) might empty the executor,
before this thread has a chance to do something with the newly enqueued tasks.
If shutdown is called from another thread, this method returns
and throws errors::shutdown_exception.
This method is thread safe.
Might throw std::system_error if one of the underlying synchronization primitives throws.
Throws errors::shutdown_exception if shutdown was called before.
*/
size_t wait_for_tasks_for(size_t count, std::chrono::milliseconds max_waiting_time);
/*
This method returns when max_count or more tasks are available for execution
or timeout_time is reached.
Returns the number of tasks available for execution when the method returns.
This method should be used as a hint, as other threads
(calling loop, for example) might empty the executor,
before this thread has a chance to do something with the newly enqueued tasks.
If shutdown is called from another thread, this method returns
and throws errors::shutdown_exception.
This method is thread safe.
Might throw std::system_error if one of the underlying synchronization primitives throws.
Throws errors::shutdown_exception if shutdown was called before.
*/
template<class clock_type, class duration_type>
size_t wait_for_tasks_until(size_t count, std::chrono::time_point<clock_type, duration_type> timeout_time);
};
Result objects
Asynchronous values and exceptions can be consumed using concurrencpp result objects. The result type represents the asynchronous result of an eager task while lazy_result represents the deferred result of a lazy task.
When a task (eager or lazy) completes, it either returns a valid value or throws an exception. In either case, this asynchronous result is marshaled to the consumer of the result object.
result objects form asymmetric coroutines - the execution of a caller-coroutine is not effected by the execution of a callee-coroutine, both coroutines can run independently. Only when consuming the result of the callee-coroutine, the caller-coroutine might be suspended awaiting the callee to complete. Up until that point both coroutines run independently. The callee-coroutine runs whether its result is consumed or not.
lazy_result objects form symmetric coroutines - execution of a callee-coroutine happens only after the suspension of the caller-coroutine. When awaiting a lazy result, the current coroutine is suspended and the lazy task associated with the lazy result starts to run. After the callee-coroutine completes and yields a result, the caller-coroutine is resumed. If a lazy result is not consumed, its associated lazy task never starts to run.
All result objects are a move-only type, and as such, they cannot be used after their content was moved to another result object. In this case, the result object is considered to be empty and attempts to call any method other than operator bool and operator = will throw.
After the asynchronous result has been pulled out of the result object (for example, by calling get or operator co_await), the result object becomes empty. Emptiness can be tested with operator bool.
Awaiting a result means to suspend the current coroutine until the result object is ready. If a valid value was returned from the associated task, it is returned from the result object. If the associated task throws an exception, it is re-thrown. At the moment of awaiting, if the result is already ready, the current coroutine resumes immediately. Otherwise, it is resumed by the thread that sets the asynchronous result or exception.
Resolving a result is similar to awaiting it. The difference is that the co_await expression will return the result object itself, in a non empty form, in a ready state. The asynchronous result can then be pulled by using get or co_await.
Every result object has a status indicating the state of the asynchronous result. The result status varies from result_status::idle (the asynchronous result or exception haven't been produced yet) to result_status::value (the associated task terminated gracefully by returning a valid value) to result_status::exception (the task terminated by throwing an exception). The status can be queried by calling (lazy_)result::status.
result type
The result type represents the result of an ongoing, asynchronous task, similar to std::future.
Aside from awaiting and resolving result-objects, they can also be waited for by calling any of result::wait, result::wait_for, result::wait_until or result::get. Waiting for a result to finish is a blocking operation (in the case the asynchronous result is not ready), and will suspend the entire thread of execution waiting for the asynchronous result to become available. Waiting operations are generally discouraged and only allowed in root-level tasks or in contexts which allow it, like blocking the main thread waiting for the rest of the application to finish gracefully, or using concurrencpp::blocking_executor or concurrencpp::thread_executor.
Awaiting result objects by using co_await (and by doing so, turning the current function/task into a coroutine as well) is the preferred way of consuming result objects, as it does not block underlying threads.
result API
class result{
/*
Creates an empty result that isn't associated with any task.
*/
result() noexcept = default;
/*
Destroys the result. Associated tasks are not cancelled.
The destructor does not block waiting for the asynchronous result to become ready.
*/
~result() noexcept = default;
/*
Moves the content of rhs to *this. After this call, rhs is empty.
*/
result(result&& rhs) noexcept = default;
/*
Moves the content of rhs to *this. After this call, rhs is empty. Returns *this.
*/
result& operator = (result&& rhs) noexcept = default;
/*
Returns true if this is a non-empty result.
Applications must not use this object if this->operator bool() is false.
*/
explicit operator bool() const noexcept;
/*
Queries the status of *this.
The returned value is any of result_status::idle, result_status::value or result_status::exception.
Throws errors::empty_result if *this is empty.
*/
result_status status() const;
/*
Blocks the current thread of execution until this result is ready,
when status() != result_status::idle.
Throws errors::empty_result if *this is empty.
Might throw std::bad_alloc if fails to allocate memory.
Might throw std::system_error if one of the underlying synchronization primitives throws.
*/
void wait();
/*
Blocks until this result is ready or duration has passed. Returns the status
of this result after unblocking.
Throws errors::empty_result if *this is empty.
Might throw std::bad_alloc if fails to allocate memory.
Might throw std::system_error if one of the underlying synchronization primitives throws.
*/
template<class duration_unit, class ratio>
result_status wait_for(std::chrono::duration<duration_unit, ratio> duration);
/*
Blocks until this result is ready or timeout_time has reached. Returns the status
of this result after unblocking.
Throws errors::empty_result if *this is empty.
Might throw std::bad_alloc if fails to allocate memory.
Might throw std::system_error if one of the underlying synchronization primitives throws.
*/
template< class clock, class duration >
result_status wait_until(std::chrono::time_point<clock, duration> timeout_time);
/*
Blocks the current thread of execution until this result is ready,
when status() != result_status::idle.
If the result is a valid value, it is returned, otherwise, get rethrows the asynchronous exception.
Throws errors::empty_result if *this is empty.
Might throw std::bad_alloc if fails to allocate memory.
Might throw std::system_error if one of the underlying synchronization primitives throws.
*/
type get();
/*
Returns an awaitable used to await this result.
If the result is already ready - the current coroutine resumes
immediately in the calling thread of execution.
If the result is not ready yet, the current coroutine is suspended
and resumed when the asynchronous result is ready,
by the thread which had set the asynchronous value or exception.
In either way, after resuming, if the result is a valid value, it is returned.
Otherwise, operator co_await rethrows the asynchronous exception.
Throws errors::empty_result if *this is empty.
*/
auto operator co_await();
/*
Returns an awaitable used to resolve this result.
After co_await expression finishes, *this is returned in a non-empty form, in a ready state.
Throws errors::empty_result if *this is empty.
*/
auto resolve();
};
lazy_result type
A lazy result object represents the result of a deferred lazy task.
lazy_result has the responsibility of both starting the associated lazy task and marshaling its deferred result back to its consumer. When awaited or resolved, the lazy result suspends the current coroutine and starts the associated lazy task. when the associated task completes, its asynchronous value is marshaled to the caller task, which is then resumed.
Sometimes, an API might return a lazy result, but applications need its associated task to run eagerly (without suspending the caller task). In this case, lazy tasks can be converted to eager tasks by calling run on its associated lazy result. In this case, the associated task will start to run inline, without suspending the caller task. The original lazy result is emptied and a valid result object that monitors the newly started task will be returned instead.
lazy_result API
class lazy_result {
/*
Creates an empty lazy result that isn't associated with any task.
*/
lazy_result() noexcept = default;
/*
Moves the content of rhs to *this. After this call, rhs is empty.
*/
lazy_result(lazy_result&& rhs) noexcept;
/*
Destroys the result. If not empty, the destructor destroys the associated task without resuming it.
*/
~lazy_result() noexcept;
/*
Moves the content of rhs to *this. After this call, rhs is empty. Returns *this.
If *this is not empty, then operator= destroys the associated task without resuming it.
*/
lazy_result& operator=(lazy_result&& rhs) noexcept;
/*
Returns true if this is a non-empty result.
Applications must not use this object if this->operator bool() is false.
*/
explicit operator bool() const noexcept;
/*
Queries the status of *this.
The returned value is any of result_status::idle, result_status::value or result_status::exception.
Throws errors::empty_result if *this is empty.
*/
result_status status() const;
/*
Returns an awaitable used to start the associated task and await this result.
If the result is already ready - the current coroutine resumes immediately
in the calling thread of execution.
If the result is not ready yet, the current coroutine is suspended and
resumed when the asynchronous result is ready,
by the thread which had set the asynchronous value or exception.
In either way, after resuming, if the result is a valid value, it is returned.
Otherwise, operator co_await rethrows the asynchronous exception.
Throws errors::empty_result if *this is empty.
*/
auto operator co_await();
/*
Returns an awaitable used to start the associated task and resolve this result.
If the result is already ready - the current coroutine resumes immediately
in the calling thread of execution.
If the result is not ready yet, the current coroutine is suspended and resumed
when the asynchronous result is ready, by the thread which
had set the asynchronous value or exception.
After co_await expression finishes, *this is returned in a non-empty form, in a ready state.
Throws errors::empty_result if *this is empty.
*/
auto resolve();
/*
Runs the associated task inline and returns a result object that monitors the newly started task.
After this call, *this is empty.
Throws errors::empty_result if *this is empty.
Might throw std::bad_alloc if fails to allocate memory.
*/
result<type> run();
};
Parallel coroutines
Regular eager coroutines start to run synchronously in the calling thread of execution. Execution might shift to another thread of execution if the coroutine undergoes a rescheduling, for example by awaiting an unready result object inside it. concurrencpp also provides parallel coroutines, which start to run inside a given executor, not in the invoking thread of execution. This style of scheduling coroutines is especially helpful when writing parallel algorithms, recursive algorithms and concurrent algorithms that use the fork-join model.
Every parallel coroutine must meet the following preconditions:
- Returns any of
result/null_result. - Gets
executor_tagas its first argument . - Gets any of
type*/type&/std::shared_ptr<type>, wheretypeis a concrete class ofexecutoras its second argument. - Contains any of
co_awaitorco_returnin its body.
If all the above applies, the function is a parallel coroutine: concurrencpp will start the coroutine suspended and immediately reschedule it to run in the provided executor. concurrencpp::executor_tag is a dummy placeholder to tell the concurrencpp runtime that this function is not a regular function, it needs to start running inside the given executor. Applications can then consume the result of the parallel coroutine by using the returned result object.
Parallel Fibonacci example:
#include "concurrencpp/concurrencpp.h"
#include <iostream>
using namespace concurrencpp;
int fibonacci_sync(int i) {
if (i == 0) {
return 0;
}
if (i == 1) {
return 1;
}
return fibonacci_sync(i - 1) + fibonacci_sync(i - 2);
}
result<int> fibonacci(executor_tag, std::shared_ptr<thread_pool_executor> tpe, const int curr) {
if (curr <= 10) {
co_return fibonacci_sync(curr);
}
auto fib_1 = fibonacci({}, tpe, curr - 1);
auto fib_2 = fibonacci({}, tpe, curr - 2);
co_return co_await fib_1 + co_await fib_2;
}
int main() {
concurrencpp::runtime runtime;
auto fibb_30 = fibonacci({}, runtime.thread_pool_executor(), 30).get();
std::cout << "fibonacci(30) = " << fibb_30 << std::endl;
return 0;
}
In this example, we calculate the 30-th member of the Fibonacci sequence in a parallel manner. We start launching each Fibonacci step in its own parallel coroutine. The first argument is a dummy executor_tag and the second argument is the threadpool executor. Every recursive step invokes a new parallel coroutine that runs in parallel. Each result is co_returned to its parent task and acquired by using co_await.
When we deem the input to be small enough to be calculated synchronously (when curr <= 10), we stop executing each recursive step in its own task and just solve the algorithm synchronously.
To compare, this is how the same code is written without using parallel coroutines, and relying on executor::submit alone. Since fibonacci returns a result<int>, submitting it recursively via executor::submit will result a result<result<int>>.
#include "concurrencpp/concurrencpp.h"
#include <iostream>
using namespace concurrencpp;
int fibonacci_sync(int i) {
if (i == 0) {
return 0;
}
if (i == 1) {
return 1;
}
return fibonacci_sync(i - 1) + fibonacci_sync(i - 2);
}
result<int> fibonacci(std::shared_ptr<thread_pool_executor> tpe, const int curr) {
if (curr <= 10) {
co_return fibonacci_sync(curr);
}
auto fib_1 = tpe->submit(fibonacci, tpe, curr - 1);
auto fib_2 = tpe->submit(fibonacci, tpe, curr - 2);
co_return co_await co_await fib_1 +
co_await co_await fib_2;
}
int main() {
concurrencpp::runtime runtime;
auto fibb_30 = fibonacci(runtime.thread_pool_executor(), 30).get();
std::cout << "fibonacci(30) = " << fibb_30 << std::endl;
return 0;
}
Result-promises
Result objects are the main way to pass data between tasks in concurrencpp and we've seen how executors and coroutines produce such objects. Sometimes we want to use the capabilities of result objects with non-tasks, for example when using a third-party library. In this case, we can complete a result object by using a result_promise. result_promise resembles a std::promise object - applications can manually set the asynchronous result or exception and make the associated result object become ready.
Just like result objects, result-promises are a move only type that becomes empty after move. Similarly, after setting a result or an exception, the result promise becomes empty as well. If a result-promise gets out of scope and no result/exception has been set, the result-promise destructor sets a concurrencpp::errors::broken_task exception using the set_exception method. Suspended and blocked tasks waiting for the associated result object are resumed/unblocked.
Result promises can convert callback style of code into async/await style of code: whenever a component requires a callback to marshal the asynchronous result, we can pass a callback that calls set_result or set_exception (depending on the asynchronous result itself) on the passed result promise, and return the associated result.
result_promise API
template <class type>
class result_promise {
/*
Constructs a valid result_promise.
Might throw std::bad_alloc if fails to allocate memory.
*/
result_promise();
/*
Moves the content of rhs to *this. After this call, rhs is empty.
*/
result_promise(result_promise&& rhs) noexcept;
/*
Destroys *this, possibly setting an errors::broken_task exception
by calling set_exception if *this is not empty at the time of destruction.
*/
~result_promise() noexcept;
/*
Moves the content of rhs to *this. After this call, rhs is empty.
*/
result_promise& operator = (result_promise&& rhs) noexcept;
/*
Returns true if this is a non-empty result-promise.
Applications must not use this object if this->operator bool() is false.
*/
explicit operator bool() const noexcept;
/*
Sets a value by constructing <<type>> from arguments... in-place.
Makes the associated result object become ready - tasks waiting for it
to become ready are unblocked.
Suspended tasks are resumed inline.
After this call, *this becomes empty.
Throws errors::empty_result_promise exception If *this is empty.
Might throw any exception that the constructor
of type(std::forward<argument_types>(arguments)...) throws.
*/
template<class ... argument_types>
void set_result(argument_types&& ... arguments);
/*
Sets an exception.
Makes the associated result object become ready - tasks waiting for it
to become ready are unblocked.
Suspended tasks are resumed inline.
After this call, *this becomes empty.
Throws errors::empty_result_promise exception If *this is empty.
Throws std::invalid_argument exception if exception_ptr is null.
*/
void set_exception(std::exception_ptr exception_ptr);
/*
A convenience method that invokes a callable with arguments... and calls set_result
with the result of the invocation.
If an exception is thrown, the thrown exception is caught and set instead by calling set_exception.
After this call, *this becomes empty.
Throws errors::empty_result_promise exception If *this is empty.
Might throw any exception that callable(std::forward<argument_types>(arguments)...)
or the contructor of type(type&&) throw.
*/
template<class callable_type, class ... argument_types>
void set_from_function(callable_type&& callable, argument_types&& ... arguments);
/*
Gets the associated result object.
Throws errors::empty_result_promise exception If *this is empty.
Throws errors::result_already_retrieved exception if this method had been called before.
*/
result<type> get_result();
};
Example: Marshaling asynchronous result using result_promise:
#include "concurrencpp/concurrencpp.h"
#include <iostream>
int main() {
concurrencpp::result_promise<std::string> promise;
auto result = promise.get_result();
std::thread my_3_party_executor([promise = std::move(promise)] () mutable {
std::this_thread::sleep_for(std::chrono::seconds(1)); //Imitate real work
promise.set_result("hello world");
});
auto asynchronous_string = result.get();
std::cout << "result promise returned string: " << asynchronous_string << std::endl;
my_3_party_executor.join();
}
In this example, We use std::thread as a third-party executor. This represents a scenario when a non-concurrencpp executor is used as part of the application life-cycle. We extract the result object before we pass the promise and block the main thread until the result becomes ready. In my_3_party_executor, we set a result as if we co_returned it.
Shared result objects
Shared results are a special kind of result objects that allow multiple consumers to access the asynchronous result, similar to std::shared_future. Different consumers from different threads can call functions like await, get and resolve in a thread safe manner.
Shared results are built from regular result objects and unlike regular result objects, they are both copyable and movable. As such, shared_result behaves like an std::shared_ptr object. If the shared result was moved to another instance, the shared result is empty, and trying to access it will throw an exception.
In order to support multiple consumers, the shared-result object will return a reference to asynchronous value instead of moving it (like a regular result object). For example, a shared_result<int>will return an int& when get,await etc. are called. If the underlying type of the shared_result is void or a reference type (like int&), they are returned as usual. If the asynchronous result is a thrown-exception, it is re-thrown.
Do note that while acquiring the asynchronous result using shared_result from multiple threads is thread-safe, the actual value might not be. For example, multiple threads can acquire an asynchronous integer by receiving its reference (int&). It does not make the integer itself thread safe. It is alright to mutate the asynchronous value if the asynchronous value is already thread safe. Alternatively, applications are encouraged to use const types to begin with (like const int), and acquire constant-references (like const int&) that prevent mutation.
shared_result API
class share_result {
/*
Creates an empty shared-result that isn't associated with any task.
*/
shared_result() noexcept = default;
/*
Destroys the shared-result. Associated tasks are not cancelled.
The destructor does not block waiting for the asynchronous result to become ready.
*/
~shared_result() noexcept = default;
/*
Converts a regular result object to a shared-result object.
After this call, rhs is empty.
Might throw std::bad_alloc if fails to allocate memory.
*/
shared_result(result<type> rhs);
/*
Copy constructor. Creates a copy of the shared result object that monitors the same task.
*/
shared_result(const shared_result&) noexcept = default;
/*
Move constructor. Moves rhs to *this. After this call, rhs is empty.
*/
shared_result(shared_result&& rhs) noexcept = default;
/*
Copy assignment operator. Copies rhs to *this and monitors the same task that rhs monitors.
*/
shared_result& operator=(const shared_result& rhs) noexcept;
/*
Move assignment operator. Moves rhs to *this. After this call, rhs is empty.
*/
shared_result& operator=(shared_result&& rhs) noexcept;
/*
Returns true if this is a non-empty shared-result.
Applications must not use this object if this->operator bool() is false.
*/
explicit operator bool() const noexcept;
/*
Queries the status of *this.
The return value is any of result_status::idle, result_status::value or result_status::exception.
Throws errors::empty_result if *this is empty.
*/
result_status status() const;
/*
Blocks the current thread of execution until this shared-result is ready,
when status() != result_status::idle.
Throws errors::empty_result if *this is empty.
Might throw std::system_error if one of the underlying synchronization primitives throws.
*/
void wait();
/*
Blocks until this shared-result is ready or duration has passed.
Returns the status of this shared-result after unblocking.
Throws errors::empty_result if *this is empty.
Might throw std::system_error if one of the underlying synchronization primitives throws.
*/
template<class duration_type, class ratio_type>
result_status wait_for(std::chrono::duration<duration_type, ratio_type> duration);
/*
Blocks until this shared-result is ready or timeout_time has reached.
Returns the status of this result after unblocking.
Throws errors::empty_result if *this is empty.
Might throw std::system_error if one of the underlying synchronization primitives throws.
*/
template<class clock_type, class duration_type>
result_status wait_until(std::chrono::time_point<clock_type, duration_type> timeout_time);
/*
Blocks the current thread of execution until this shared-result is ready,
when status() != result_status::idle.
If the result is a valid value, a reference to it is returned,
otherwise, get rethrows the asynchronous exception.
Throws errors::empty_result if *this is empty.
Might throw std::system_error if one of the underlying synchronization primitives throws.
*/
std::add_lvalue_reference_t<type> get();
/*
Returns an awaitable used to await this shared-result.
If the shared-result is already ready - the current coroutine resumes
immediately in the calling thread of execution.
If the shared-result is not ready yet, the current coroutine is
suspended and resumed when the asynchronous result is ready,
by the thread which had set the asynchronous value or exception.
In either way, after resuming, if the result is a valid value, a reference to it is returned.
Otherwise, operator co_await rethrows the asynchronous exception.
Throws errors::empty_result if *this is empty.
*/
auto operator co_await();
/*
Returns an awaitable used to resolve this shared-result.
After co_await expression finishes, *this is returned in a non-empty form, in a ready state.
Throws errors::empty_result if *this is empty.
*/
auto resolve();
};
shared_result example
#include "concurrencpp/concurrencpp.h"
#include <iostream>
#include <chrono>
concurrencpp::result<void> consume_shared_result(concurrencpp::shared_result<int> shared_result,
std::shared_ptr<concurrencpp::executor> resume_executor) {
std::cout << "Awaiting shared_result to have a value" << std::endl;
const auto& async_value = co_await shared_result;
concurrencpp::resume_on(resume_executor);
std::cout << "In thread id " << std::this_thread::get_id() << ", got: " << async_value << ", memory address: " << &async_value << std::endl;
}
int main() {
concurrencpp::runtime runtime;
auto result = runtime.background_executor()->submit([] {
std::this_thread::sleep_for(std::chrono::seconds(1));
return 100;
});
concurrencpp::shared_result<int> shared_result(std::move(result));
concurrencpp::result<void> results[8];
for (size_t i = 0; i < 8; i++) {
results[i] = consume_shared_result(shared_result, runtime.thread_pool_executor());
}
std::cout << "Main thread waiting for all consumers to finish" << std::endl;
auto tpe = runtime.thread_pool_executor();
auto all_consumed = concurrencpp::when_all(tpe, std::begin(results), std::end(results)).run();
all_consumed.get();
std::cout << "All consumers are done, exiting" << std::endl;
return 0;
}
Termination in concurrencpp
When the runtime object gets out of scope of main, the application terminates. The runtime iterates each stored executor and calls its shutdown method. Trying to access either the timer-queue or any executor throws errors::runtime_shutdown exception. When an executor shuts down, it clears its inner task queues, destroying un-executed task objects. If a task object stores a concurrencpp-coroutine, that coroutine is resumed inline and an errors::broken_task exception is thrown. In any case where a runtime_shutdown or a broken_task exception is thrown, applications should terminate their current code-flow gracefully as soon as possible. Those exceptions should not be ignored.
Resume executors
Many concurrencpp asynchronous actions will require an executor as their resume executor. When an asynchronous action (implemented as a coroutine) can finish synchronously, it resumes immediately in the calling thread of execution. If the asynchronous action can't finish synchronously, it will be resumed when it finishes, inside the given resume-executor. For example, when_any utility function requires a resume-executor as its first argument. when_any returns a lazy_result which becomes ready when at least one given result becomes ready. If one of the results is already ready at the moment of calling when_any, the calling coroutine is resumed synchronously in the calling thread of execution. If not, the calling coroutine will be resumed when at least of result is finished, inside the given resume-executor. Resume executors are important because they mandate where coroutines are resumed in cases where it's not clear where a coroutine is supposed to be resumed (for example, in the case of when_any and when_all), or in cases where the asynchronous action is processed inside one of the concurrencpp workers, which are only used to process that specific action, and not application code.
Utility functions
make_ready_result function
make_ready_result creates a ready result object from given arguments. Awaiting such result will cause the current coroutine to resume immediately. get and operator co_await will return the constructed value.
/*
Creates a ready result object by building <<type>> from arguments&&... in-place.
Might throw any exception that the constructor
of type(std::forward<argument_types>(arguments)...) throws.
Might throw std::bad_alloc exception if fails to allocate memory.
*/
template<class type, class ... argument_types>
result<type> make_ready_result(argument_types&& ... arguments);
/*
An overload for void type.
Might throw std::bad_alloc exception if fails to allocate memory.
*/
result<void> make_ready_result();
make_exceptional_result function
make_exceptional_result creates a ready result object from a given exception. Awaiting such result will cause the current coroutine to resume immediately. get and operator co_await will re-throw the given exception.
/*
Creates a ready result object from an exception pointer.
The returned result object will re-throw exception_ptr when calling get or await.
Throws std::invalid_argument if exception_ptr is null.
Might throw std::bad_alloc exception if fails to allocate memory.
*/
template<class type>
result<type> make_exceptional_result(std::exception_ptr exception_ptr);
/*
Overload. Similar to make_exceptional_result(std::exception_ptr),
but gets an exception object directly.
Might throw any exception that the constructor of exception_type(std::move(exception)) might throw.
Might throw std::bad_alloc exception if fails to allocate memory.
*/
template<class type, class exception_type>
result<type> make_exceptional_result(exception_type exception);
when_all function
when_all is a utility function that creates a lazy result object which becomes ready when all input results are completed. Awaiting this lazy result returns all input-result objects in a ready state, ready to be consumed.
when_all function comes with three flavors - one that accepts a heterogeneous range of result objects, another that gets a pair of iterators to a range of result objects of the same type, and lastly an overload that accepts no results objects at all. In the case of no input result objects - the function returns a ready result object of an empty tuple.
If one of the passed result-objects is empty, an exception will be thrown. In this case, input-result objects are unaffected by the function and can be used again after the exception was handled. If all input result objects are valid, they are emptied by this function, and returned in a valid and ready state as the output result.
Currently, when_all only accepts result objects.
All overloads accept a resume executor as their first parameter. When awaiting a result returned by when_all, the caller coroutine will be resumed by the given resume executor.
/*
Creates a result object that becomes ready when all the input results become ready.
Passed result objects are emptied and returned as a tuple.
Throws std::invalid_argument if any of the passed result objects is empty.
Might throw an std::bad_alloc exception if no memory is available.
*/
template<class ... result_types>
lazy_result<std::tuple<typename std::decay<result_types>::type...>>
when_all(std::shared_ptr<executor_type> resume_executor,
result_types&& ... results);
/*
Overload. Similar to when_all(result_types&& ...) but receives a pair of iterators referencing a range.
Passed result objects are emptied and returned as a vector.
If begin == end, the function returns immediately with an empty vector.
Throws std::invalid_argument if any of the passed result objects is empty.
Might throw an std::bad_alloc exception if no memory is available.
*/
template<class iterator_type>
lazy_result<std::vector<typename std::iterator_traits<iterator_type>::value_type>>
when_all(std::shared_ptr<executor_type> resume_executor,
iterator_type begin, iterator_type end);
/*
Overload. Returns a ready result object that doesn't monitor any asynchronous result.
Might throw an std::bad_alloc exception if no memory is available.
*/
lazy_result<std::tuple<>> when_all(std::shared_ptr<executor_type> resume_executor);
when_any function
when_any is a utility function that creates a lazy result object which becomes ready when at least one input result is completed. Awaiting this result will return a helper struct containing all input-result objects plus the index of the completed task. It could be that by the time of consuming the ready result, other results might have already completed asynchronously. Applications can call when_any repeatedly in order to consume ready results as they complete until all results are consumed.
when_any function comes with only two flavors - one that accepts a heterogeneous range of result objects and another that gets a pair of iterators to a range of result-objects of the same type. Unlike when_all, there is no meaning in awaiting at least one task to finish when the range of results is completely empty. Hence, there is no overload with no arguments. Also, the overload of two iterators will throw an exception if those iterators reference an empty range (when begin == end).
If one of the passed result-objects is empty, an exception will be thrown. In any case an exception is thrown, input-result objects are unaffected by the function and can be used again after the exception was handled. If all input result objects are valid, they are emptied by this function, and returned in a valid state as the output result.
Currently, when_any only accepts result objects.
All overloads accept a resume executor as their first parameter. When awaiting a result returned by when_any, the caller coroutine will be resumed by the given resume executor.
/*
Helper struct returned from when_any.
index is the position of the ready result in results sequence.
results is either an std::tuple or an std::vector of the results that were passed to when_any.
*/
template <class sequence_type>
struct when_any_result {
std::size_t index;
sequence_type results;
};
/*
Creates a result object that becomes ready when at least one of the input results is ready.
Passed result objects are emptied and returned as a tuple.
Throws std::invalid_argument if any of the passed result objects is empty.
Might throw an std::bad_alloc exception if no memory is available.
*/
template<class ... result_types>
lazy_result<when_any_result<std::tuple<result_types...>>>
when_any(std::shared_ptr<executor_type> resume_executor,
result_types&& ... results);
/*
Overload. Similar to when_any(result_types&& ...) but receives a pair of iterators referencing a range.
Passed result objects are emptied and returned as a vector.
Throws std::invalid_argument if begin == end.
Throws std::invalid_argument if any of the passed result objects is empty.
Might throw an std::bad_alloc exception if no memory is available.
*/
template<class iterator_type>
lazy_result<when_any_result<std::vector<typename std::iterator_traits<iterator_type>::value_type>>>
when_any(std::shared_ptr<executor_type> resume_executor,
iterator_type begin, iterator_type end);
resume_on function
resume_on returns an awaitable that suspends the current coroutine and resumes it inside given executor. This is an important function that makes sure a coroutine is running in the right executor. For example, applications might schedule a background task using the background_executor and await the returned result object. In this case, the awaiting coroutine will be resumed inside the background executor. A call to resume_on with another cpu-bound executor makes sure that cpu-bound lines of code will not run on the background executor once the background task is completed. If a coroutine was re-scheduled to run on another executor using resume_on, but that executor is shut down before it can resume it, that coroutine is resumed and an erros::broken_task exception is thrown. In this case, applications need to quite gracefully.
/*
Returns an awaitable that suspends the current coroutine and resumes it inside executor.
Might throw any exception that executor_type::enqueue throws.
*/
template<class executor_type>
auto resume_on(std::shared_ptr<executor_type> executor);
Timers and Timer queues
concurrencpp also provides timers and timer queues. Timers are objects that define asynchronous actions running on an executor within a well-defined interval of time. There are three types of timers - regular timers, onshot-timers and delay objects.
Regular timers have four properties that define them:
- Callable - a callable that will be scheduled to run as a task periodically.
- Executor - an executor that schedules the callable to run periodically.
- Due time - from the time of creation, the interval in milliseconds the timer will be scheduled to run for the first time.
- Frequency - from the time the timer was scheduled to run for the first time, the interval in milliseconds the callable will be scheduled to run periodically, until the timer is destructed or cancelled.
Like other objects in concurrencpp, timers are a move only type that can be empty. When a timer is destructed or timer::cancel is called, the timer cancels its scheduled but not yet executed tasks. Ongoing tasks are uneffected. The timer callable must be thread safe. It is recommended to set the due time and the frequency of timers to a granularity of 50 milliseconds.
A timer queue is a concurrencpp worker that manages a collection of timers and processes them in just one thread of execution. It is also the agent used to create new timers. When a timer deadline (whether it is the timer's due-time or frequency) has reached, the timer queue "fires" the timer by scheduling its callable to run on the associated executor as a task.
Just like executors, timer queues also adhere to the RAII concept. When the runtime object gets out of scope, It shuts down the timer queue, cancelling all pending timers. After a timer queue has been shut down, any subsequent call to make_timer, make_onshot_timer and make_delay_object will throw an errors::runtime_shutdown exception. Applications must not try to shut down timer queues by themselves.
timer_queue API:
class timer_queue {
/*
Destroys this timer_queue.
*/
~timer_queue() noexcept;
/*
Shuts down this timer_queue:
Tells the underlying thread of execution to quit and joins it.
Cancels all pending timers.
After this call, invocation of any method besides shutdown
and shutdown_requested will throw an errors::runtime_shutdown.
If shutdown had been called before, this method has no effect.
*/
void shutdown() noexcept;
/*
Returns true if shutdown had been called before, false otherwise.
*/
bool shutdown_requested() const noexcept;
/*
Creates a new running timer where *this is the associated timer_queue.
Throws std::invalid_argument if executor is null.
Throws errors::runtime_shutdown if shutdown had been called before.
Might throw std::bad_alloc if fails to allocate memory.
Might throw std::system_error if the one of the underlying synchronization primitives throws.
*/
template<class callable_type, class ... argumet_types>
timer make_timer(
std::chrono::milliseconds due_time,
std::chrono::milliseconds frequency,
std::shared_ptr<concurrencpp::executor> executor,
callable_type&& callable,
argumet_types&& ... arguments);
/*
Creates a new one-shot timer where *this is the associated timer_queue.
Throws std::invalid_argument if executor is null.
Throws errors::runtime_shutdown if shutdown had been called before.
Might throw std::bad_alloc if fails to allocate memory.
Might throw std::system_error if the one of the underlying synchronization primitives throws.
*/
template<class callable_type, class ... argumet_types>
timer make_one_shot_timer(
std::chrono::milliseconds due_time,
std::shared_ptr<concurrencpp::executor> executor,
callable_type&& callable,
argumet_types&& ... arguments);
/*
Creates a new delay object where *this is the associated timer_queue.
Throws std::invalid_argument if executor is null.
Throws errors::runtime_shutdown if shutdown had been called before.
Might throw std::bad_alloc if fails to allocate memory.
Might throw std::system_error if the one of the underlying synchronization primitives throws.
*/
result<void> make_delay_object(
std::chrono::milliseconds due_time,
std::shared_ptr<concurrencpp::executor> executor);
};
timer API:
class timer {
/*
Creates an empty timer.
*/
timer() noexcept = default;
/*
Cancels the timer, if not empty.
*/
~timer() noexcept;
/*
Moves the content of rhs to *this.
rhs is empty after this call.
*/
timer(timer&& rhs) noexcept = default;
/*
Moves the content of rhs to *this.
rhs is empty after this call.
Returns *this.
*/
timer& operator = (timer&& rhs) noexcept;
/*
Cancels this timer.
After this call, the associated timer_queue will not schedule *this
to run again and *this becomes empty.
Scheduled, but not yet executed tasks are cancelled.
Ongoing tasks are uneffected.
This method has no effect if *this is empty or the associated timer_queue has already expired.
Might throw std::system_error if one of the underlying synchronization primitives throws.
*/
void cancel();
/*
Returns the associated executor of this timer.
Throws concurrencpp::errors::empty_timer is *this is empty.
*/
std::shared_ptr<executor> get_executor() const;
/*
Returns the associated timer_queue of this timer.
Throws concurrencpp::errors::empty_timer is *this is empty.
*/
std::weak_ptr<timer_queue> get_timer_queue() const;
/*
Returns the due time of this timer.
Throws concurrencpp::errors::empty_timer is *this is empty.
*/
std::chrono::milliseconds get_due_time() const;
/*
Returns the frequency of this timer.
Throws concurrencpp::errors::empty_timer is *this is empty.
*/
std::chrono::milliseconds get_frequency() const;
/*
Sets new frequency for this timer.
Callables already scheduled to run at the time of invocation are not affected.
Throws concurrencpp::errors::empty_timer is *this is empty.
*/
void set_frequency(std::chrono::milliseconds new_frequency);
/*
Returns true is *this is not an empty timer, false otherwise.
The timer should not be used if this->operator bool() is false.
*/
explicit operator bool() const noexcept;
};
Regular timer example:
#include "concurrencpp/concurrencpp.h"
#include <iostream>
using namespace std::chrono_literals;
int main() {
concurrencpp::runtime runtime;
std::atomic_size_t counter = 1;
concurrencpp::timer timer = runtime.timer_queue()->make_timer(
1500ms,
2000ms,
runtime.thread_pool_executor(),
[&] {
const auto c = counter.fetch_add(1);
std::cout << "timer was invoked for the " << c << "th time" << std::endl;
});
std::this_thread::sleep_for(12s);
return 0;
}
In this example we create a regular timer by using the timer queue. The timer schedules its callable after 1.5 seconds, then fires its callable every 2 seconds. The given callable runs in the threadpool executor.
Oneshot timers
A oneshot timer is a one-time timer with only a due time - after it schedules its callable to run once it never reschedules it to run again.
Oneshot timer example:
#include "concurrencpp/concurrencpp.h"
#include <iostream>
using namespace std::chrono_literals;
int main() {
concurrencpp::runtime runtime;
concurrencpp::timer timer = runtime.timer_queue()->make_one_shot_timer(
3000ms,
runtime.thread_executor(),
[&] {
std::cout << "hello and goodbye" << std::endl;
});
std::this_thread::sleep_for(4s);
return 0;
}
In this example, we create a timer that runs only once - after 3 seconds from its creation, the timer will schedule to run its callable on a new thread of execution (using concurrencpp::thread_executor).
Delay objects
A delay object is a result object that becomes ready when its due time is reached. Applications can co_await this result object to delay the current coroutine in a non-blocking way. The current coroutine is resumed by the executor that was passed to make_delay_object.
Delay object example:
#include "concurrencpp/concurrencpp.h"
#include <iostream>
using namespace std::chrono_literals;
concurrencpp::null_result delayed_task(
std::shared_ptr<concurrencpp::timer_queue> tq,
std::shared_ptr<concurrencpp::thread_pool_executor> ex) {
size_t counter = 1;
while(true) {
std::cout << "task was invoked " << counter << " times." << std::endl;
counter++;
co_await tq->make_delay_object(1500ms, ex);
}
}
int main() {
concurrencpp::runtime runtime;
delayed_task(runtime.timer_queue(), runtime.thread_pool_executor());
std::this_thread::sleep_for(10s);
return 0;
}
In this example, we created a coroutine (that does not marshal any result or thrown exception), which delays itself in a loop by calling co_await on a delay object.
Generators
A generator is a lazy, synchronous coroutine that is able to produce a stream of values to consume. Generators use the co_yield keyword to yield values back to their consumers.
Example:
A generator that yields the n-th member of the Sequence S(n) = 1 + 2 + 3 + ... + n where n <= 100:
concurrencpp::generator<int> sequence() {
int i = 1;
int sum = 0;
while (i <= 100) {
sum += i;
++i;
co_yield sum;
}
}
int main() {
for (auto value : sequence()) {
std::cout << value << std::end;
}
return 0;
}
Generators are meant to be used synchronously - they can only use the co_yield keyword and must not use the co_await keyword. A generator will continue to produce values as long as the co_yield keyword is called. If the co_return keyword is called (explicitly or implicitly), then the generator will stop producing values. Similarly, if an exception is thrown then the generator will stop producing values and the thrown exception will be re-thrown to the consumer of the generator.
Generators are meant to be used in a range-for loop: Generators implicitly produce two iterators - begin and end which control the execution of the for loop. These iterators should not be handled or accessed manually.
When a generator is created, it starts as a lazy task. When its begin method is called, the generator is resumed for the first time and an iterator is returned. The lazy task is resumed repeatedly by calling operator++ on the returned iterator. The returned iterator will be equal to end iterator when the generator finishes execution either by exiting gracefully or throwing an exception. As mentioned earlier, this happens behind the scenes by the inner mechanism of the loop and the generator, and should not be called directly.
Like other objects in concurrencpp, Generators are a move-only type. After a generator was moved, it is considered empty and trying to access its inner methods (other than operator bool) will throw an exception. The emptiness of a generator should not generally occur - it is advised to consume generators upon their creation in a for loop and not to try to call its methods individually.
generator API
class generator {
/*
Move constructor. After this call, rhs is empty.
*/
generator(generator&& rhs) noexcept;
/*
Destructor. Invalidates existing iterators.
*/
~generator() noexcept;
generator(const generator& rhs) = delete;
generator& operator=(generator&& rhs) = delete;
generator& operator=(const generator& rhs) = delete;
/*
Returns true if this generator is not empty.
Applications must not use this object if this->operator bool() is false.
*/
explicit operator bool() const noexcept;
/*
Starts running this generator and returns an iterator.
Throws errors::empty_generator if *this is empty.
Re-throws any exception that is thrown inside the generator code.
*/
iterator begin();
/*
Returns an end iterator.
*/
static generator_end_iterator end() noexcept;
};
class generator_iterator {
using value_type = std::remove_reference_t<type>;
using reference = value_type&;
using pointer = value_type*;
using iterator_category = std::input_iterator_tag;
using difference_type = std::ptrdiff_t;
/*
Resumes the suspended generator and returns *this.
Re-throws any exception that was thrown inside the generator code.
*/
generator_iterator& operator++();
/*
Post-increment version of operator++.
*/
void operator++(int);
/*
Returns the latest value produced by the associated generator.
*/
reference operator*() const noexcept;
/*
Returns a pointer to the latest value produced by the associated generator.
*/
pointer operator->() const noexcept;
/*
Comparision operators.
*/
friend bool operator==(const generator_iterator& it0, const generator_iterator& it1) noexcept;
friend bool operator==(const generator_iterator& it, generator_end_iterator) noexcept;
friend bool operator==(generator_end_iterator end_it, const generator_iterator& it) noexcept;
friend bool operator!=(const generator_iterator& it, generator_end_iterator end_it) noexcept;
friend bool operator!=(generator_end_iterator end_it, const generator_iterator& it) noexcept;
};
The runtime object
The concurrencpp runtime object is the agent used to acquire, store and create new executors.
The runtime must be created as a value type as soon as the main function starts to run. When the concurrencpp runtime gets out of scope, it iterates over its stored executors and shuts them down one by one by calling executor::shutdown. Executors then exit their inner work loop and any subsequent attempt to schedule a new task will throw a concurrencpp::runtime_shutdown exception. The runtime also contains the global timer queue used to create timers and delay objects. Upon destruction, stored executors will destroy unexecuted tasks, and wait for ongoing tasks to finish. If an ongoing task tries to use an executor to spawn new tasks or schedule its own task continuation - an exception will be thrown. In this case, ongoing tasks need to quit as soon as possible, allowing their underlying executors to quit. The timer queue will also be shut down, cancelling all running timers. With this RAII style of code, no tasks can be processed before the creation of the runtime object, and while/after the runtime gets out of scope. This frees concurrent applications from needing to communicate termination messages explicitly. Tasks are free use executors as long as the runtime object is alive.
runtime API
class runtime {
/*
Creates a runtime object with default options.
*/
runtime();
/*
Creates a runtime object with user defined options.
*/
runtime(const concurrencpp::runtime_options& options);
/*
Destroys this runtime object.
Calls executor::shutdown on each monitored executor.
Calls timer_queue::shutdown on the global timer queue.
*/
~runtime() noexcept;
/*
Returns this runtime timer queue used to create new times.
*/
std::shared_ptr<concurrencpp::timer_queue> timer_queue() const noexcept;
/*
Returns this runtime concurrencpp::inline_executor
*/
std::shared_ptr<concurrencpp::inline_executor> inline_executor() const noexcept;
/*
Returns this runtime concurrencpp::thread_pool_executor
*/
std::shared_ptr<concurrencpp::thread_pool_executor> thread_pool_executor() const noexcept;
/*
Returns this runtime concurrencpp::background_executor
*/
std::shared_ptr<concurrencpp::thread_pool_executor> background_executor() const noexcept;
/*
Returns this runtime concurrencpp::thread_executor
*/
std::shared_ptr<concurrencpp::thread_executor> thread_executor() const noexcept;
/*
Creates a new concurrencpp::worker_thread_executor and registers it in this runtime.
Might throw std::bad_alloc or std::system_error if any underlying memory or system resource could not have been acquired.
*/
std::shared_ptr<concurrencpp::worker_thread_executor> make_worker_thread_executor();
/*
Creates a new concurrencpp::manual_executor and registers it in this runtime.
Might throw std::bad_alloc or std::system_error if any underlying memory or system resource could not have been acquired.
*/
std::shared_ptr<concurrencpp::manual_executor> make_manual_executor();
/*
Creates a new user defined executor and registers it in this runtime.
executor_type must be a valid concrete class of concurrencpp::executor.
Might throw std::bad_alloc if no memory is available.
Might throw any exception that the constructor of <<executor_type>> might throw.
*/
template<class executor_type, class ... argument_types>
std::shared_ptr<executor_type> make_executor(argument_types&& ... arguments);
/*
returns the version of concurrencpp that the library was built with.
*/
static std::tuple<unsigned int, unsigned int, unsigned int> version() noexcept;
};
Creating user-defined executors
As mentioned before, Applications can create their own custom executor type by inheriting the derivable_executor class. There are a few points to consider when implementing user defined executors: The most important thing is to remember that executors are used from multiple threads, so implemented methods must be thread-safe.
New executors can be created using runtime::make_executor. Applications must not create new executors with plain instantiation (such as std::make_shared or plain new), only by using runtime::make_executor. Also, applications must not try to re-instantiate the built-in concurrencpp executors, like the thread_pool_executor or the thread_executor, those executors must only be accessed through their existing instance in the runtime object.
Another important point is to handle shutdown correctly: shutdown, shutdown_requested and enqueue should all monitor the executor state and behave accordingly when invoked:
shutdownshould tell underlying threads to quit and then join them.shutdownmight be called multiple times, and the method must handle this scenario by ignoring any subsequent call toshutdownafter the first invocation.enqueuemust throw aconcurrencpp::errors::runtime_shutdownexception ifshutdownhad been called before.
task objects
Implementing executors is one of the rare cases applications need to work with concurrencpp::task class directly. concurrencpp::task is a std::function like object, but with a few differences. Like std::function, the task object stores a callable that acts as the asynchronous operation. Unlike std::function, task is a move only type. On invocation, task objects receive no parameters and return void. Moreover, every task object can be invoked only once. After the first invocation, the task object becomes empty. Invoking an empty task object is equivalent to invoking an empty lambda ([]{}), and will not throw any exception. Task objects receive their callable as a forwarding reference (type&& where type is a template parameter), and not by copy (like std::function). Construction of the stored callable happens in-place. This allows task objects to contain callables that are move-only type (like std::unique_ptr and concurrencpp::result). Task objects try to use different methods to optimize the usage of the stored types, for example, task objects apply the short-buffer-optimization (sbo) for regular, small callables, and will inline calls to std::coroutine_handle<void> by calling them directly without virtual dispatch.
task API
class task {
/*
Creates an empty task object.
*/
task() noexcept;
/*
Creates a task object by moving the stored callable of rhs to *this.
If rhs is empty, then *this will also be empty after construction.
After this call, rhs is empty.
*/
task(task&& rhs) noexcept;
/*
Creates a task object by storing callable in *this.
<<typename std::decay<callable_type>::type>> will be in-place-
constructed inside *this by perfect forwarding callable.
*/
template<class callable_type>
task(callable_type&& callable);
/*
Destroys stored callable, does nothing if empty.
*/
~task() noexcept;
/*
If *this is empty, does nothing.
Invokes stored callable, and immediately destroys it.
After this call, *this is empty.
May throw any exception that the invoked callable may throw.
*/
void operator()();
/*
Moves the stored callable of rhs to *this.
If rhs is empty, then *this will also be empty after this call.
If *this already contains a stored callable, operator = destroys it first.
*/
task& operator=(task&& rhs) noexcept;
/*
If *this is not empty, task::clear destroys the stored callable and empties *this.
If *this is empty, clear does nothing.
*/
void clear() noexcept;
/*
Returns true if *this stores a callable. false otherwise.
*/
explicit operator bool() const noexcept;
/*
Returns true if *this stores a callable,
and that stored callable has the same type as <<typename std::decay<callable_type>::type>>
*/
template<class callable_type>
bool contains() const noexcept;
};
When implementing user-defined executors, it is up to the implementation to store tasks (when enqueue is called), and execute them according to the executor inner-mechanism.
Example: using a user-defined executor:
#include "concurrencpp/concurrencpp.h"
#include <iostream>
#include <queue>
#include <thread>
#include <mutex>
#include <condition_variable>
class logging_executor : public concurrencpp::derivable_executor<logging_executor> {
private:
mutable std::mutex _lock;
std::queue<concurrencpp::task> _queue;
std::condition_variable _condition;
bool _shutdown_requested;
std::thread _thread;
const std::string _prefix;
void work_loop() {
while (true) {
std::unique_lock<std::mutex> lock(_lock);
if (_shutdown_requested) {
return;
}
if (!_queue.empty()) {
auto task = std::move(_queue.front());
_queue.pop();
lock.unlock();
std::cout << _prefix << " A task is being executed" << std::endl;
task();
continue;
}
_condition.wait(lock, [this] {
return !_queue.empty() || _shutdown_requested;
});
}
}
public:
logging_executor(std::string_view prefix) :
derivable_executor<logging_executor>("logging_executor"),
_shutdown_requested(false),
_prefix(prefix) {
_thread = std::thread([this] {
work_loop();
});
}
void enqueue(concurrencpp::task task) override {
std::cout << _prefix << " A task is being enqueued!" << std::endl;
std::unique_lock<std::mutex> lock(_lock);
if (_shutdown_requested) {
throw concurrencpp::errors::runtime_shutdown("logging executor - executor was shutdown.");
}
_queue.emplace(std::move(task));
_condition.notify_one();
}
void enqueue(std::span<concurrencpp::task> tasks) override {
std::cout << _prefix << tasks.size() << " tasks are being enqueued!" << std::endl;
std::unique_lock<std::mutex> lock(_lock);
if (_shutdown_requested) {
throw concurrencpp::errors::runtime_shutdown("logging executor - executor was shutdown.");
}
for (auto& task : tasks) {
_queue.emplace(std::move(task));
}
_condition.notify_one();
}
int max_concurrency_level() const noexcept override {
return 1;
}
bool shutdown_requested() const noexcept override {
std::unique_lock<std::mutex> lock(_lock);
return _shutdown_requested;
}
void shutdown() noexcept override {
std::cout << _prefix << " shutdown requested" << std::endl;
std::unique_lock<std::mutex> lock(_lock);
if (_shutdown_requested) return; //nothing to do.
_shutdown_requested = true;
lock.unlock();
_condition.notify_one();
_thread.join();
}
};
int main() {
concurrencpp::runtime runtime;
auto logging_ex = runtime.make_executor<logging_executor>("Session #1234");
for (size_t i = 0; i < 10; i++) {
logging_ex->post([] {
std::cout << "hello world" << std::endl;
});
}
std::getchar();
return 0;
}
In this example, we created an executor which logs actions like enqueuing a task or executing it. We implement the executor interface, and we request the runtime to create and store an instance of it by calling runtime::make_executor. The rest of the application behaves exactly the same as if we were to use non user-defined executors.
Supported platforms and tools
- Operating systems: Linux, macOS, Windows (Windows 10 and above)
- Compilers: MSVC (Visual Studio 2019 version 16.8.2 and above), Clang (Clang-11 and above)
- Tools: CMake (3.16 and above)
Building, installing and testing
Building the library on Windows (release mode)
$ git clone https://github.com/David-Haim/concurrencpp.git
$ cd concurrencpp
$ cmake -S . -B build/lib
$ cmake --build build/lib --config Release
Running the tests on Windows (debug + release mode)
$ git clone https://github.com/David-Haim/concurrencpp.git
$ cd concurrencpp
$ cmake -S test -B build/test
$ cmake --build build/test
<# for release mode: cmake --build build/test --config Release #>
$ cd build/test
$ ctest . -V -C Debug
<# for release mode: ctest . -V -C Release #>
Building the library on *nix platforms (release mode)
$ git clone https://github.com/David-Haim/concurrencpp.git
$ cd concurrencpp
$ cmake -DCMAKE_BUILD_TYPE=Release -S . -B build/lib
$ cmake --build build/lib
#optional, install the library: sudo cmake --install build/lib
Running the tests on *nix platforms
With clang, it is also possible to run the tests with TSAN (thread sanitizer) support.
$ git clone https://github.com/David-Haim/concurrencpp.git
$ cd concurrencpp
$ cmake -S test -B build/test
#for release mode: cmake -DCMAKE_BUILD_TYPE=Release -S test -B build/test
#for TSAN mode: cmake -DCMAKE_BUILD_TYPE=Release -DENABLE_THREAD_SANITIZER=Yes -S test -B build/test
$ cmake --build build/test
$ cd build/test
$ ctest . -V
Via vcpkg on Windows and *nix platforms
Alternatively to building and installing the library manually, developers may get stable releases of concurrencpp as vcpkg packages:
$ vcpkg install concurrencpp
Experimenting with the built-in sandbox
concurrencpp comes with a built-in sandbox program which developers can modify and experiment, without having to install or link the compiled library to a different code-base. In order to play with the sandbox, developers can modify sandbox/main.cpp and compile the application using the following commands:
Building and running the sandbox on Windows:
$ cmake -S sandbox -B build/sandbox
$ cmake --build build/sandbox
<# for release mode: cmake --build build/sandbox --config Release #>
$ ./build/sandbox <# runs the sandbox>
Building and running the sandbox on *nix platforms:
$ cmake -S sandbox -B build/sandbox
#for release mode: cmake -DCMAKE_BUILD_TYPE=Release -S sandbox -B build/sandbox
$ cmake --build build/sandbox
$ ./build/sandbox #runs the sandboxAuthor: David-Haim
Source Code: https://github.com/David-Haim/concurrencpp
License: MIT License
1655268300
The C Beginner's Handbook: Learn C Programming Language basics in just a few hours
This C Beginner's Handbook follows the 80/20 rule. You'll learn 80% of the C programming language in 20% of the time.
This approach will give you a well-rounded overview of the language.
This handbook does not try to cover everything under the sun related to C. It focuses on the core of the language, trying to simplify the more complex topics.
And note: You can get a PDF and ePub version of this C Beginner's Handbook here.
Enjoy!
Introduction to C
C is probably the most widely known programming language. It is used as the reference language for computer science courses all over the world, and it's probably the language that people learn the most in school along with Python and Java.
I remember it being my second programming language ever, after Pascal.
C is not just what students use to learn programming. It's not an academic language. And I would say it's not the easiest language, because C is a rather low level programming language.
Today, C is widely used in embedded devices, and it powers most of the Internet servers, which are built using Linux. The Linux kernel is built using C, and this also means that C powers the core of all Android devices. We can say that C code runs a good portion of the entire world. Right now. Pretty remarkable.
When it was created, C was considered a high level language, because it was portable across machines. Today we kind of take for granted that we can run a program written on a Mac on Windows or Linux, perhaps using Node.js or Python.
Once upon a time, this was not the case at all. What C brought to the table was a language that was simple to implement and that had a compiler that could be easily ported to different machines.
I said compiler: C is a compiled programming language, like Go, Java, Swift or Rust. Other popular programming language like Python, Ruby or JavaScript are interpreted. The difference is consistent: a compiled language generates a binary file that can be directly executed and distributed.
C is not garbage collected. This means we have to manage memory ourselves. It's a complex task and one that requires a lot of attention to prevent bugs, but it is also what makes C ideal to write programs for embedded devices like Arduino.
C does not hide the complexity and the capabilities of the machine underneath. You have a lot of power, once you know what you can do.
I want to introduce the first C program now, which we'll call "Hello, World!"
hello.c
#include <stdio.h>
int main(void) {
printf("Hello, World!");
}
Let's describe the program source code: we first import the stdio library (the name stands for standard input-output library).
This library gives us access to input/output functions.
C is a very small language at its core, and anything that's not part of the core is provided by libraries. Some of those libraries are built by normal programmers, and made available for others to use. Some other libraries are built into the compiler. Like stdio and others.
stdio is the library that provides the printf() function.
This function is wrapped into a main() function. The main() function is the entry point of any C program.
But what is a function, anyway?
A function is a routine that takes one or more arguments, and returns a single value.
In the case of main(), the function gets no arguments, and returns an integer. We identify that using the void keyword for the argument, and the int keyword for the return value.
The function has a body, which is wrapped in curly braces. Inside the body we have all the code that the function needs to perform its operations.
The printf() function is written differently, as you can see. It has no return value defined, and we pass a string, wrapped in double quotes. We didn't specify the type of the argument.
That's because this is a function invocation. Somewhere, inside the stdio library, printf is defined as
int printf(const char *format, ...);
You don't need to understand what this means now, but in short, this is the definition. And when we call printf("Hello, World!");, that's where the function is run.
The main() function we defined above:
#include <stdio.h>
int main(void) {
printf("Hello, World!");
}
will be run by the operating system when the program is executed.
How do we execute a C program?
As mentioned, C is a compiled language. To run the program we must first compile it. Any Linux or macOS computer already comes with a C compiler built-in. For Windows, you can use the Windows Subsystem for Linux (WSL).
In any case, when you open the terminal window you can type gcc, and this command should return an error saying that you didn't specify any file:
That's good. It means the C compiler is there, and we can start using it.
Now type the program above into a hello.c file. You can use any editor, but for the sake of simplicity I'm going to use the nano editor in the command line:
Type the program:
Now press ctrl-X to exit:
Confirm by pressing the y key, then press enter to confirm the file name:
That's it, we should be back to the terminal now:
Now type
gcc hello.c -o hello
The program should give you no errors:
but it should have generated a hello executable. Now type
./hello
to run it:
I prepend ./ to the program name to tell the terminal that the command is in the current folder
Awesome!
Now if you call ls -al hello, you can see that the program is only 12KB in size:
This is one of the pros of C: it's highly optimized, and this is also one of the reasons it's this good for embedded devices that have a very limited amount of resources.
Variables and types
C is a statically typed language.
This means that any variable has an associated type, and this type is known at compilation time.
This is very different than how you work with variables in Python, JavaScript, PHP and other interpreted languages.
When you create a variable in C, you have to specify the type of a variable at the declaration.
In this example we initialize a variable age with type int:
int age;
A variable name can contain any uppercase or lowercase letter, can contain digits and the underscore character, but it can't start with a digit. AGE and Age10 are valid variable names, 1age is not.
You can also initialize a variable at declaration, specifying the initial value:
int age = 37;
Once you declare a variable, you are then able to use it in your program code. You can change its value at any time, using the = operator for example, like in age = 100; (provided the new value is of the same type).
In this case:
#include <stdio.h>
int main(void) {
int age = 0;
age = 37.2;
printf("%u", age);
}
the compiler will raise a warning at compile time, and will convert the decimal number to an integer value.
The C built-in data types are int, char, short, long, float, double, long double. Let's find out more about those.
Integer numbers
C provides us the following types to define integer values:
charintshortlong
Most of the time, you'll likely use an int to store an integer. But in some cases, you might want to choose one of the other 3 options.
The char type is commonly used to store letters of the ASCII chart, but it can be used to hold small integers from -128 to 127. It takes at least 1 byte.
int takes at least 2 bytes. short takes at least 2 bytes. long takes at least 4 bytes.
As you can see, we are not guaranteed the same values for different environments. We only have an indication. The problem is that the exact numbers that can be stored in each data type depends on the implementation and the architecture.
We're guaranteed that short is not longer than int. And we're guaranteed long is not shorter than int.
The ANSI C spec standard determines the minimum values of each type, and thanks to it we can at least know what's the minimum value we can expect to have at our disposal.
If you are programming C on an Arduino, different board will have different limits.
On an Arduino Uno board, int stores a 2 byte value, ranging from -32,768 to 32,767. On a Arduino MKR 1010, int stores a 4 bytes value, ranging from -2,147,483,648 to 2,147,483,647. Quite a big difference.
On all Arduino boards, short stores a 2 bytes value, ranging from -32,768 to 32,767. long store 4 bytes, ranging from -2,147,483,648 to 2,147,483,647.
Unsigned integers
For all the above data types, we can prepend unsigned to start the range at 0, instead of a negative number. This might make sense in many cases.
unsigned charwill range from0to at least255unsigned intwill range from0to at least65,535unsigned shortwill range from0to at least65,535unsigned longwill range from0to at least4,294,967,295
The problem with overflow
Given all those limits, a question might come up: how can we make sure our numbers do not exceed the limit? And what happens if we do exceed the limit?
If you have an unsigned int number at 255, and you increment it, you'll get 256 in return. As expected. If you have an unsigned char number at 255, and you increment it, you'll get 0 in return. It resets starting from the initial possible value.
If you have a unsigned char number at 255 and you add 10 to it, you'll get the number 9:
#include <stdio.h>
int main(void) {
unsigned char j = 255;
j = j + 10;
printf("%u", j); /* 9 */
}
If you don't have a signed value, the behavior is undefined. It will basically give you a huge number which can vary, like in this case:
#include <stdio.h>
int main(void) {
char j = 127;
j = j + 10;
printf("%u", j); /* 4294967177 */
}
In other words, C does not protect you from going over the limits of a type. You need to take care of this yourself.
Warnings when declaring the wrong type
When you declare the variable and initialize it with the wrong value, the gcc compiler (the one you're probably using) should warn you:
#include <stdio.h>
int main(void) {
char j = 1000;
}
hello.c:4:11: warning: implicit conversion
from 'int' to
'char' changes value from 1000 to -24
[-Wconstant-conversion]
char j = 1000;
~ ^~~~
1 warning generated.
And it also warns you in direct assignments:
#include <stdio.h>
int main(void) {
char j;
j = 1000;
}
But not if you increase the number using, for example, +=:
#include <stdio.h>
int main(void) {
char j = 0;
j += 1000;
}
Floating point numbers
Floating point types can represent a much larger set of values than integers can, and can also represent fractions, something that integers can't do.
Using floating point numbers, we represent numbers as decimal numbers times powers of 10.
You might see floating point numbers written as
1.29e-3-2.3e+5
and in other seemingly weird ways.
The following types:
floatdoublelong double
are used to represent numbers with decimal points (floating point types). All can represent both positive and negative numbers.
The minimum requirements for any C implementation is that float can represent a range between 10^-37 and 10^+37, and is typically implemented using 32 bits. double can represent a bigger set of numbers. long double can hold even more numbers.
The exact figures, as with integer values, depend on the implementation.
On a modern Mac, a float is represented in 32 bits, and has a precision of 24 significant bits. 8 bits are used to encode the exponent.
A double number is represented in 64 bits, with a precision of 53 significant bits. 11 bits are used to encode the exponent.
The type long double is represented in 80 bits, has a precision of 64 significant bits. 15 bits are used to encode the exponent.
On your specific computer, how can you determine the specific size of the types? You can write a program to do that:
#include <stdio.h>
int main(void) {
printf("char size: %lu bytes\n", sizeof(char));
printf("int size: %lu bytes\n", sizeof(int));
printf("short size: %lu bytes\n", sizeof(short));
printf("long size: %lu bytes\n", sizeof(long));
printf("float size: %lu bytes\n", sizeof(float));
printf("double size: %lu bytes\n",
sizeof(double));
printf("long double size: %lu bytes\n",
sizeof(long double));
}
In my system, a modern Mac, it prints:
char size: 1 bytes
int size: 4 bytes
short size: 2 bytes
long size: 8 bytes
float size: 4 bytes
double size: 8 bytes
long double size: 16 bytes
Constants
Let's now talk about constants.
A constant is declared similarly to variables, except it is prepended with the const keyword, and you always need to specify a value.
Like this:
const int age = 37;
This is perfectly valid C, although it is common to declare constants uppercase, like this:
const int AGE = 37;
It's just a convention, but one that can greatly help you while reading or writing a C program as it improves readability. Uppercase name means constant, lowercase name means variable.
A constant name follows the same rules for variable names: can contain any uppercase or lowercase letter, can contain digits and the underscore character, but it can't start with a digit. AGE and Age10 are valid variable names, 1AGE is not.
Another way to define constants is by using this syntax:
#define AGE 37
In this case, you don't need to add a type, and you don't also need the = equal sign, and you omit the semicolon at the end.
The C compiler will infer the type from the value specified, at compile time.
Operators
C offers us a wide variety of operators that we can use to operate on data.
In particular, we can identify various groups of operators:
- arithmetic operators
- comparison operators
- logical operators
- compound assignment operators
- bitwise operators
- pointer operators
- structure operators
- miscellaneous operators
In this section I'm going to detail all of them, using 2 imaginary variables a and b as examples.
I am keeping bitwise operators, structure operators and pointer operators out of this list, to keep things simpler
Arithmetic operators
In this macro group I am going to separate binary operators and unary operators.
Binary operators work using two operands:
| OPERATOR | NAME | EXAMPLE |
|---|---|---|
= | Assignment | a = b |
+ | Addition | a + b |
- | Subtraction | a - b |
* | Multiplication | a * b |
/ | Division | a / b |
% | Modulo | a % b |
Unary operators only take one operand:
| OPERATOR | NAME | EXAMPLE |
|---|---|---|
+ | Unary plus | +a |
- | Unary minus | -a |
++ | Increment | a++ or ++a |
-- | Decrement | a-- or --a |
The difference between a++ and ++a is that a++ increments the a variable after using it. ++a increments the a variable before using it.
For example:
int a = 2;
int b;
b = a++ /* b is 2, a is 3 */
b = ++a /* b is 4, a is 4 */
The same applies to the decrement operator.
Comparison operators
| OPERATOR | NAME | EXAMPLE |
|---|---|---|
== | Equal operator | a == b |
!= | Not equal operator | a != b |
> | Bigger than | a > b |
< | Less than | a < b |
>= | Bigger than or equal to | a >= b |
<= | Less than or equal to | a <= b |
Logical operators
!NOT (example:!a)&&AND (example:a && b)||OR (example:a || b)
Those operators are great when working with boolean values.
Compound assignment operators
Those operators are useful to perform an assignment and at the same time perform an arithmetic operation:
| OPERATOR | NAME | EXAMPLE |
|---|---|---|
+= | Addition assignment | a += b |
-= | Subtraction assignment | a -= b |
*= | Multiplication assignment | a *= b |
/= | Division assignment | a /= b |
%= | Modulo assignment | a %= b |
The ternary operator
The ternary operator is the only operator in C that works with 3 operands, and it’s a short way to express conditionals.
This is how it looks:
<condition> ? <expression> : <expression>
Example:
a ? b : c
If a is evaluated to true, then the b statement is executed, otherwise c is.
The ternary operator is functionality-wise same as an if/else conditional, except it is shorter to express and it can be inlined into an expression.
sizeof
The sizeof operator returns the size of the operand you pass. You can pass a variable, or even a type.
Example usage:
#include <stdio.h>
int main(void) {
int age = 37;
printf("%ld\n", sizeof(age));
printf("%ld", sizeof(int));
}
Operator precedence
With all those operators (and more, which I haven't covered in this post, including bitwise, structure operators, and pointer operators), we must pay attention when using them together in a single expression.
Suppose we have this operation:
int a = 2;
int b = 4;
int c = b + a * a / b - a;
What's the value of c? Do we get the addition being executed before the multiplication and the division?
There is a set of rules that help us solve this puzzle.
In order from less precedence to more precedence, we have:
- the
=assignment operator - the
+and-binary operators - the
*and/operators - the
+and-unary operators
Operators also have an associativity rule, which is always left to right except for the unary operators and the assignment.
In:
int c = b + a * a / b - a;
We first execute a * a / b, which, due to being left-to-right, we can separate into a * a and the result / b: 2 * 2 = 4, 4 / 4 = 1.
Then we can perform the sum and the subtraction: 4 + 1 - 2. The value of c is 3.
In all cases, however, I want to make sure you realize you can use parentheses to make any similar expression easier to read and comprehend.
Parentheses have higher priority over anything else.
The above example expression can be rewritten as:
int c = b + ((a * a) / b) - a;
and we don't have to think about it that much.
Conditionals
Any programming language provides the programmers the ability to perform choices.
We want to do X in some cases, and Y in other cases.
We want to check data, and make choices based on the state of that data.
C provides us 2 ways to do so.
The first is the if statement, with its else helper, and the second is the switch statement.
if
In an if statement, you can check for a condition to be true, and then execute the block provided in the curly brackets:
int a = 1;
if (a == 1) {
/* do something */
}
You can append an else block to execute a different block if the original condition turns out to be false:
int a = 1;
if (a == 2) {
/* do something */
} else {
/* do something else */
}
Beware of one common source of bugs - always use the comparison operator == in comparisons, and not the assignment operator =. If you don't, the if conditional check will always be true, unless the argument is 0, for example if you do:
int a = 0;
if (a = 0) {
/* never invoked */
}
Why does this happen? Because the conditional check will look for a boolean result (the result of a comparison), and the 0 number always equates to a false value. Everything else is true, including negative numbers.
You can have multiple else blocks by stacking together multiple if statements:
int a = 1;
if (a == 2) {
/* do something */
} else if (a == 1) {
/* do something else */
} else {
/* do something else again */
}
switch
If you need to do too many if / else / if blocks to perform a check, perhaps because you need to check the exact value of a variable, then switch can be very useful to you.
You can provide a variable as condition, and a series of case entry points for each value you expect:
int a = 1;
switch (a) {
case 0:
/* do something */
break;
case 1:
/* do something else */
break;
case 2:
/* do something else */
break;
}
We need a break keyword at the end of each case to avoid the next case being executed when the one before ends. This "cascade" effect can be useful in some creative ways.
You can add a "catch-all" case at the end, labeled default:
int a = 1;
switch (a) {
case 0:
/* do something */
break;
case 1:
/* do something else */
break;
case 2:
/* do something else */
break;
default:
/* handle all the other cases */
break;
}
Loops
C offers us three ways to perform a loop: for loops, while loops and do while loops. They all allow you to iterate over arrays, but with a few differences. Let's see them in detail.
For loops
The first and probably most common way to perform a loop is for loops.
Using the for keyword we can define the rules of the loop up front, and then provide the block that is going to be executed repeatedly.
Like this:
for (int i = 0; i <= 10; i++) {
/* instructions to be repeated */
}
The (int i = 0; i <= 10; i++) block contains 3 parts of the looping details:
- the initial condition (
int i = 0) - the test (
i <= 10) - the increment (
i++)
We first define a loop variable, in this case named i. i is a common variable name to be used for loops, along with j for nested loops (a loop inside another loop). It's just a convention.
The variable is initialized at the 0 value, and the first iteration is done. Then it is incremented as the increment part says (i++ in this case, incrementing by 1), and all the cycle repeats until you get to the number 10.
Inside the loop main block we can access the variable i to know at which iteration we are. This program should print 0 1 2 3 4 5 5 6 7 8 9 10:
for (int i = 0; i <= 10; i++) {
/* instructions to be repeated */
printf("%u ", i);
}
Loops can also start from a high number, and go a lower number, like this:
for (int i = 10; i > 0; i--) {
/* instructions to be repeated */
}
You can also increment the loop variable by 2 or another value:
for (int i = 0; i < 1000; i = i + 30) {
/* instructions to be repeated */
}
While loops
While loops is simpler to write than a for loop, because it requires a bit more work on your part.
Instead of defining all the loop data up front when you start the loop, like you do in the for loop, using while you just check for a condition:
while (i < 10) {
}
This assumes that i is already defined and initialized with a value.
And this loop will be an infinite loop unless you increment the i variable at some point inside the loop. An infinite loop is bad because it will block the program, allowing nothing else to happen.
This is what you need for a "correct" while loop:
int i = 0;
while (i < 10) {
/* do something */
i++;
}
There's one exception to this, and we'll see it in one minute. Before, let me introduce do while.
Do while loops
While loops are great, but there might be times when you need to do one particular thing: you want to always execute a block, and then maybe repeat it.
This is done using the do while keyword. In a way it's very similar to a while loop, but slightly different:
int i = 0;
do {
/* do something */
i++;
} while (i < 10);
The block that contains the /* do something */ comment is always executed at least once, regardless of the condition check at the bottom.
Then, until i is less than 10, we'll repeat the block.
Breaking out of a loop using break
In all the C loops we have a way to break out of a loop at any point in time, immediately, regardless of the conditions set for the loop.
This is done using the break keyword.
This is useful in many cases. You might want to check for the value of a variable, for example:
for (int i = 0; i <= 10; i++) {
if (i == 4 && someVariable == 10) {
break;
}
}
Having this option to break out of a loop is particularly interesting for while loops (and do while too), because we can create seemingly infinite loops that end when a condition occurs. You define this inside the loop block:
int i = 0;
while (1) {
/* do something */
i++;
if (i == 10) break;
}
It's rather common to have this kind of loop in C.
Arrays
An array is a variable that stores multiple values.
Every value in the array, in C, must have the same type. This means you will have arrays of int values, arrays of double values, and more.
You can define an array of int values like this:
int prices[5];
You must always specify the size of the array. C does not provide dynamic arrays out of the box (you have to use a data structure like a linked list for that).
You can use a constant to define the size:
const int SIZE = 5;
int prices[SIZE];
You can initialize an array at definition time, like this:
int prices[5] = { 1, 2, 3, 4, 5 };
But you can also assign a value after the definition, in this way:
int prices[5];
prices[0] = 1;
prices[1] = 2;
prices[2] = 3;
prices[3] = 4;
prices[4] = 5;
Or, more practical, using a loop:
int prices[5];
for (int i = 0; i < 5; i++) {
prices[i] = i + 1;
}
And you can reference an item in the array by using square brackets after the array variable name, adding an integer to determine the index value. Like this:
prices[0]; /* array item value: 1 */
prices[1]; /* array item value: 2 */
Array indexes start from 0, so an array with 5 items, like the prices array above, will have items ranging from prices[0] to prices[4].
The interesting thing about C arrays is that all elements of an array are stored sequentially, one right after another. Not something that normally happens with higher-level programming languages.
Another interesting thing is this: the variable name of the array, prices in the above example, is a pointer to the first element of the array. As such it can be used like a normal pointer.
More on pointers soon.
Strings
In C, strings are one special kind of array: a string is an array of char values:
char name[7];
I introduced the char type when I introduced types, but in short it is commonly used to store letters of the ASCII chart.
A string can be initialized like you initialize a normal array:
char name[7] = { "F", "l", "a", "v", "i", "o" };
Or more conveniently with a string literal (also called string constant), a sequence of characters enclosed in double quotes:
char name[7] = "Flavio";
You can print a string via printf() using %s:
printf("%s", name);
Do you notice how "Flavio" is 6 chars long, but I defined an array of length 7? Why? This is because the last character in a string must be a 0 value, the string terminator, and we must make space for it.
This is important to keep in mind especially when manipulating strings.
Speaking of manipulating strings, there's one important standard library that is provided by C: string.h.
This library is essential because it abstracts many of the low level details of working with strings, and provides us with a set of useful functions.
You can load the library in your program by adding on top:
#include <string.h>
And once you do that, you have access to:
strcpy()to copy a string over another stringstrcat()to append a string to another stringstrcmp()to compare two strings for equalitystrncmp()to compare the firstncharacters of two stringsstrlen()to calculate the length of a string
and many, many more.
Pointers
Pointers are one of the most confusing/challenging parts of C, in my opinion. Especially if you are new to programming, but also if you come from a higher level programming language like Python or JavaScript.
In this section I want to introduce them in the simplest yet not-dumbed-down way possible.
A pointer is the address of a block of memory that contains a variable.
When you declare an integer number like this:
int age = 37;
We can use the & operator to get the value of the address in memory of a variable:
printf("%p", &age); /* 0x7ffeef7dcb9c */
I used the %p format specified in printf() to print the address value.
We can assign the address to a variable:
int *address = &age;
Using int *address in the declaration, we are not declaring an integer variable, but rather a pointer to an integer.
We can use the pointer operator * to get the value of the variable an address is pointing to:
int age = 37;
int *address = &age;
printf("%u", *address); /* 37 */
This time we are using the pointer operator again, but since it's not a declaration this time it means "the value of the variable this pointer points to".
In this example we declare an age variable, and we use a pointer to initialize the value:
int age;
int *address = &age;
*address = 37;
printf("%u", *address);
When working with C, you'll find that a lot of things are built on top of this simple concept. So make sure you familiarize with it a bit by running the above examples on your own.
Pointers are a great opportunity because they force us to think about memory addresses and how data is organized.
Arrays are one example. When you declare an array:
int prices[3] = { 5, 4, 3 };
The prices variable is actually a pointer to the first item of the array. You can get the value of the first item using this printf() function in this case:
printf("%u", *prices); /* 5 */
The cool thing is that we can get the second item by adding 1 to the prices pointer:
printf("%u", *(prices + 1)); /* 4 */
And so on for all the other values.
We can also do many nice string manipulation operations, since strings are arrays under the hood.
We also have many more applications, including passing the reference of an object or a function around to avoid consuming more resources to copy it.
Functions
Functions are the way we can structure our code into subroutines that we can:
- give a name to
- call when we need them
Starting from your very first program, a "Hello, World!", you immediately make use of C functions:
#include <stdio.h>
int main(void) {
printf("Hello, World!");
}
The main() function is a very important function, as it's the entry point for a C program.
Here's another function:
void doSomething(int value) {
printf("%u", value);
}
Functions have 4 important aspects:
- they have a name, so we can invoke ("call") them later
- they specify a return value
- they can have arguments
- they have a body, wrapped in curly braces
The function body is the set of instructions that are executed any time we invoke a function.
If the function has no return value, you can use the keyword void before the function name. Otherwise you specify the function return value type (int for an integer, float for a floating point value, const char * for a string, etc).
You cannot return more than one value from a function.
A function can have arguments. They are optional. If it does not have them, inside the parentheses we insert void, like this:
void doSomething(void) {
/* ... */
}
In this case, when we invoke the function we'll call it with nothing in the parentheses:
doSomething();
If we have one parameter, we specify the type and the name of the parameter, like this:
void doSomething(int value) {
/* ... */
}
When we invoke the function, we'll pass that parameter in the parentheses, like this:
doSomething(3);
We can have multiple parameters, and if so we separate them using a comma, both in the declaration and in the invocation:
void doSomething(int value1, int value2) {
/* ... */
}
doSomething(3, 4);
Parameters are passed by copy. This means that if you modify value1, its value is modified locally. The value outside of the function, where it was passed in the invocation, does not change.
If you pass a pointer as a parameter, you can modify that variable value because you can now access it directly using its memory address.
You can't define a default value for a parameter. C++ can do that (and so Arduino Language programs can), but C can't.
Make sure you define the function before calling it, or the compiler will raise a warning and an error:
➜ ~ gcc hello.c -o hello; ./hello
hello.c:13:3: warning: implicit declaration of
function 'doSomething' is invalid in C99
[-Wimplicit-function-declaration]
doSomething(3, 4);
^
hello.c:17:6: error: conflicting types for
'doSomething'
void doSomething(int value1, char value2) {
^
hello.c:13:3: note: previous implicit declaration
is here
doSomething(3, 4);
^
1 warning and 1 error generated.
The warning you get regards the ordering, which I already mentioned.
The error is about another thing, related. Since C does not "see" the function declaration before the invocation, it must make assumptions. And it assumes the function to return int. The function however returns void, hence the error.
If you change the function definition to:
int doSomething(int value1, int value2) {
printf("%d %d\n", value1, value2);
return 1;
}
you'd just get the warning, and not the error:
➜ ~ gcc hello.c -o hello; ./hello
hello.c:14:3: warning: implicit declaration of
function 'doSomething' is invalid in C99
[-Wimplicit-function-declaration]
doSomething(3, 4);
^
1 warning generated.
In any case, make sure you declare the function before using it. Either move the function up, or add the function prototype in a header file.
Inside a function, you can declare variables.
void doSomething(int value) {
int doubleValue = value * 2;
}
A variable is created at the point of invocation of the function and is destroyed when the function ends. It's not visible from the outside.
Inside a function, you can call the function itself. This is called recursion and it's something that offers peculiar opportunities.
Input and output
C is a small language, and the "core" of C does not include any Input/Output (I/O) functionality.
This is not something unique to C, of course. It's common for the language core to be agnostic of I/O.
In the case of C, Input/Output is provided to us by the C Standard Library via a set of functions defined in the stdio.h header file.
You can import this library using:
#include <stdio.h>
on top of your C file.
This library provides us with, among many other functions:
printf()scanf()sscanf()fgets()fprintf()
Before describing what those functions do, I want to take a minute to talk about I/O streams.
We have 3 kinds of I/O streams in C:
stdin(standard input)stdout(standard output)stderr(standard error)
With I/O functions we always work with streams. A stream is a high level interface that can represent a device or a file. From the C standpoint, we don't have any difference in reading from a file or reading from the command line: it's an I/O stream in any case.
That's one thing to keep in mind.
Some functions are designed to work with a specific stream, like printf(), which we use to print characters to stdout. Using its more general counterpart fprintf(), we can specify which stream to write to.
Since I started talking about printf(), let's introduce it now.
printf() is one of the first functions you'll use when learning C programming.
In its simplest usage form, you pass it a string literal:
printf("hey!");
and the program will print the content of the string to the screen.
You can print the value of a variable. But it's a bit tricky because you need to add a special character, a placeholder, which changes depending on the type of the variable. For example we use %d for a signed decimal integer digit:
int age = 37;
printf("My age is %d", age);
We can print more than one variable by using commas:
int age_yesterday = 37;
int age_today = 36;
printf("Yesterday my age was %d and today is %d", age_yesterday, age_today);
There are other format specifiers like %d:
%cfor a char%sfor a char%ffor floating point numbers%pfor pointers
and many more.
We can use escape characters in printf(), like \n which we can use to make the output create a new line.
scanf()
printf() is used as an output function. I want to introduce an input function now, so we can say we can do all the I/O thing: scanf().
This function is used to get a value from the user running the program, from the command line.
We must first define a variable that will hold the value we get from the input:
int age;
Then we call scanf() with 2 arguments: the format (type) of the variable, and the address of the variable:
scanf("%d", &age);
If we want to get a string as input, remember that a string name is a pointer to the first character, so you don't need the & character before it:
char name[20];
scanf("%s", name);
Here's a little program that uses both printf() and scanf():
#include <stdio.h>
int main(void) {
char name[20];
printf("Enter your name: ");
scanf("%s", name);
printf("you entered %s", name);
}
Variable scope
When you define a variable in a C program, depending on where you declare it, it will have a different scope.
This means that it will be available in some places, but not in others.
The position determines 2 types of variables:
- global variables
- local variables
This is the difference: a variable declared inside a function is a local variable, like this:
int main(void) {
int age = 37;
}
Local variables are only accessible from within the function, and when the function ends they stop their existence. They are cleared from the memory (with some exceptions).
A variable defined outside a function is a global variable, like in this example:
int age = 37;
int main(void) {
/* ... */
}
Global variables are accessible from any function of the program, and they are available for the whole execution of the program, until it ends.
I mentioned that local variables are not available any more after the function ends.
The reason is that local variables are declared on the stack, by default, unless you explicitly allocate them on the heap using pointers. But then you have to manage the memory yourself.
Static variables
Inside a function, you can initialize a static variable using the static keyword.
I said "inside a function" because global variables are static by default, so there's no need to add the keyword.
What's a static variable? A static variable is initialized to 0 if no initial value is specified, and it retains the value across function calls.
Consider this function:
int incrementAge() {
int age = 0;
age++;
return age;
}
If we call incrementAge() once, we'll get 1 as the return value. If we call it more than once, we'll always get 1 back, because age is a local variable and it's re-initialized to 0 on every single function call.
If we change the function to:
int incrementAge() {
static int age = 0;
age++;
return age;
}
Now every time we call this function, we'll get an incremented value:
printf("%d\n", incrementAge());
printf("%d\n", incrementAge());
printf("%d\n", incrementAge());
will give us
1
2
3
We can also omit initializing age to 0 in static int age = 0;, and just write static int age; because static variables are automatically set to 0 when created.
We can also have static arrays. In this case, each single item in the array is initialized to 0:
int incrementAge() {
static int ages[3];
ages[0]++;
return ages[0];
}
Global variables
In this section I want to talk more about the difference between global and local variables.
A local variable is defined inside a function, and it's only available inside that function.
Like this:
#include <stdio.h>
int main(void) {
char j = 0;
j += 10;
printf("%u", j); //10
}
j is not available anywhere outside the main function.
A global variable is defined outside of any function, like this:
#include <stdio.h>
char i = 0;
int main(void) {
i += 10;
printf("%u", i); //10
}
A global variable can be accessed by any function in the program. Access is not limited to reading the value: the variable can be updated by any function.
Due to this, global variables are one way we have of sharing the same data between functions.
The main difference with local variables is that the memory allocated for variables is freed once the function ends.
Global variables are only freed when the program ends.
Type definitions
The typedef keyword in C allows you to defined new types.
Starting from the built-in C types, we can create our own types, using this syntax:
typedef existingtype NEWTYPE
The new type we create is usually, by convention, uppercase.
This it to distinguish it more easily, and immediately recognize it as type.
For example we can define a new NUMBER type that is an int:
typedef int NUMBER
and once you do so, you can define new NUMBER variables:
NUMBER one = 1;
Now you might ask: why? Why not just use the built-in type int instead?
Well, typedef gets really useful when paired with two things: enumerated types and structures.
Enumerated types
Using the typedef and enum keywords we can define a type that can have either one value or another.
It's one of the most important uses of the typedef keyword.
This is the syntax of an enumerated type:
typedef enum {
//...values
} TYPENAME;
The enumerated type we create is usually, by convention, uppercase.
Here is a simple example:
typedef enum {
true,
false
} BOOLEAN;
C comes with a bool type, so this example is not really practical, but you get the idea.
Another example is to define weekdays:
typedef enum {
monday,
tuesday,
wednesday,
thursday,
friday,
saturday,
sunday
} WEEKDAY;
Here's a simple program that uses this enumerated type:
#include <stdio.h>
typedef enum {
monday,
tuesday,
wednesday,
thursday,
friday,
saturday,
sunday
} WEEKDAY;
int main(void) {
WEEKDAY day = monday;
if (day == monday) {
printf("It's monday!");
} else {
printf("It's not monday");
}
}
Every item in the enum definition is paired to an integer, internally. So in this example monday is 0, tuesday is 1 and so on.
This means the conditional could have been if (day == 0) instead of if (day == monday), but it's way simpler for us humans to reason with names rather than numbers, so it's a very convenient syntax.
Structures
Using the struct keyword we can create complex data structures using basic C types.
A structure is a collection of values of different types. Arrays in C are limited to a type, so structures can prove to be very interesting in a lot of use cases.
This is the syntax of a structure:
struct <structname> {
//...variables
};
Example:
struct person {
int age;
char *name;
};
You can declare variables that have as type that structure by adding them after the closing curly bracket, before the semicolon, like this:
struct person {
int age;
char *name;
} flavio;
Or multiple ones, like this:
struct person {
int age;
char *name;
} flavio, people[20];
In this case I declare a single person variable named flavio, and an array of 20 person named people.
We can also declare variables later on, using this syntax:
struct person {
int age;
char *name;
};
struct person flavio;
We can initialize a structure at declaration time:
struct person {
int age;
char *name;
};
struct person flavio = { 37, "Flavio" };
and once we have a structure defined, we can access the values in it using a dot:
struct person {
int age;
char *name;
};
struct person flavio = { 37, "Flavio" };
printf("%s, age %u", flavio.name, flavio.age);
We can also change the values using the dot syntax:
struct person {
int age;
char *name;
};
struct person flavio = { 37, "Flavio" };
flavio.age = 38;
Structures are very useful because we can pass them around as function parameters, or return values, embedding various variables within them. Each variable has a label.
It's important to note that structures are passed by copy, unless of course you pass a pointer to a struct, in which case it's passed by reference.
Using typedef we can simplify the code when working with structures.
Let's look at an example:
typedef struct {
int age;
char *name;
} PERSON;
The structure we create using typedef is usually, by convention, uppercase.
Now we can declare new PERSON variables like this:
PERSON flavio;
and we can initialize them at declaration in this way:
PERSON flavio = { 37, "Flavio" };
Command line parameters
In your C programs, you might need to accept parameters from the command line when the command launches.
For simple needs, all you need to do to do so is change the main() function signature from
int main(void)
to
int main (int argc, char *argv[])
argc is an integer number that contains the number of parameters that were provided in the command line.
argv is an array of strings.
When the program starts, we are provided the arguments in those 2 parameters.
Note that there's always at least one item in the argv array: the name of the program
Let's take the example of the C compiler we use to run our programs, like this:
gcc hello.c -o hello
If this was our program, we'd have argc being 4 and argv being an array containing
gcchello.c-ohello
Let's write a program that prints the arguments it receives:
#include <stdio.h>
int main (int argc, char *argv[]) {
for (int i = 0; i < argc; i++) {
printf("%s\n", argv[i]);
}
}
If the name of our program is hello and we run it like this: ./hello, we'd get this as output:
./hello
If we pass some random parameters, like this: ./hello a b c we'd get this output to the terminal:
./hello
a
b
c
This system works great for simple needs. For more complex needs, there are commonly used packages like getopt.
Headers files
Simple programs can be put in a single file. But when your program grows larger it's impossible to keep it all in just one file.
You can move parts of a program to a separate file. Then you create a header file.
A header file looks like a normal C file, except it ends with .h instead of .c. Instead of the implementations of your functions and the other parts of a program, it holds the declarations.
You already used header files when you first used the printf() function, or other I/O function, and you had to type:
#include <stdio.h>
to use it.
#include is a preprocessor directive.
The preprocessor goes and looks up the stdio.h file in the standard library because you used brackets around it. To include your own header files, you'll use quotes, like this:
#include "myfile.h"
The above will look up myfile.h in the current folder.
You can also use a folder structure for libraries:
#include "myfolder/myfile.h"
Let's look at an example. This program calculates the years since a given year:
#include <stdio.h>
int calculateAge(int year) {
const int CURRENT_YEAR = 2020;
return CURRENT_YEAR - year;
}
int main(void) {
printf("%u", calculateAge(1983));
}
Suppose I want to move the calculateAge function to a separate file.
I create a calculate_age.c file:
int calculateAge(int year) {
const int CURRENT_YEAR = 2020;
return CURRENT_YEAR - year;
}
And a calculate_age.h file where I put the function prototype, which is the same as the function in the .c file, except the body:
int calculateAge(int year);
Now in the main .c file we can go and remove the calculateAge() function definition, and we can import calculate_age.h, which will make the calculateAge() function available:
#include <stdio.h>
#include "calculate_age.h"
int main(void) {
printf("%u", calculateAge(1983));
}
Don't forget that to compile a program composed by multiple files, you need to list them all in the command line, like this:
gcc -o main main.c calculate_age.c
And with more complex setups, a Makefile is necessary to tell the compiler how to compile the program.
The preprocessor
The preprocessor is a tool that helps us a lot when programming with C. It is part of the C Standard, just like the language, the compiler, and the standard library.
It parses our program and makes sure that the compiler gets all the things it needs before going on with the process.
What does it do, in practice?
For example, it looks up all the header files you include with the #include directive.
It also looks at every constant you defined using #define and substitutes it with its actual value.
That's just the start. I mentioned those 2 operations because they are the most common ones. The preprocessor can do a lot more.
Did you notice #include and #define have a # at the beginning? That's common to all the preprocessor directives. If a line starts with #, that's taken care of by the preprocessor.
Conditionals
One of the things we can do is to use conditionals to change how our program will be compiled, depending on the value of an expression.
For example we can check if the DEBUG constant is 0:
#include <stdio.h>
const int DEBUG = 0;
int main(void) {
#if DEBUG == 0
printf("I am NOT debugging\n");
#else
printf("I am debugging\n");
#endif
}
Symbolic constants
We can define a symbolic constant:
#define VALUE 1
#define PI 3.14
#define NAME "Flavio"
When we use NAME or PI or VALUE in our program, the preprocessor replaces its name with the value before executing the program.
Symbolic constants are very useful because we can give names to values without creating variables at compilation time.
Macros
With #define we can also define a macro. The difference between a macro and a symbolic constant is that a macro can accept an argument and typically contains code, while a symbolic constant is a value:
#define POWER(x) ((x) * (x))
Notice the parentheses around the arguments: this is a good practice to avoid issues when the macro is replaced in the precompilation process.
Then we can use it in our code like this:
printf("%u\n", POWER(4)); //16
The big difference with functions is that macros do not specify the type of their arguments or return values, which might be handy in some cases.
Macros, however, are limited to one line definitions.
If defined
We can check if a symbolic constant or a macro is defined using #ifdef:
#include <stdio.h>
#define VALUE 1
int main(void) {
#ifdef VALUE
printf("Value is defined\n");
#else
printf("Value is not defined\n");
#endif
}
We also have #ifndev to check for the opposite (macro not defined).
We can also use #if defined and #if !defined to do the same task.
It's common to wrap some block of code into a block like this:
#if 0
#endif
to temporarily prevent it from running, or to use a DEBUG symbolic constant:
#define DEBUG 0
#if DEBUG
//code only sent to the compiler
//if DEBUG is not 0
#endif
Predefined symbolic constants you can use
The preprocessor also defines a number of symbolic constants you can use, identified by the 2 underscores before and after the name, including:
__LINE__translates to the current line in the source code file__FILE__translates to the name of the file__DATE__translates to the compilation date, in theMmm gg aaaaformat__TIME__translates to the compilation time, in thehh:mm:ssformat
Conclusion
Thanks a lot for reading this handbook!
I hope it will inspire you to know more about C.
This story was originally published at https://www.freecodecamp.org/news/the-c-beginners-handbook/