Reputation: 145
I'm trying to write milt-threaded HTTP proxy for learning C++/socket/HTTP
I'm looking for a HTTP client library like HttpURLConnection available in Java.
I looked at some libraries eg, libcurl for C/C++. These libraries can make http request but they will return with the full content. I need a library that can read content partially in a buffer so that I'm able to ship it immediately to requesting client, without storing the entire content in memory.
Any links/suggestions is highly appreciated :)
thank you!
Upvotes: 5
Views: 8333
Reputation: 4703
Microsoft released "cpprestsdk" a modern, cross-platform, asynchronous rest client / server with json serialization / deserialization support. https://github.com/microsoft/cpprestsdk
Upvotes: 1
Reputation: 5353
Beast is an open source C++ library which supports the functionality you desire. It has a "Writer" concept which supports suspending and resuming, incremental rendering of the body, and coroutines: http://vinniefalco.github.io/beast/beast/types/Writer.html
The library is here: http://vinniefalco.github.io/
Here's a complete example program:
#include <beast/http.hpp>
#include <boost/asio.hpp>
#include <iostream>
#include <string>
int main()
{
// Normal boost::asio setup
std::string const host = "boost.org";
boost::asio::io_service ios;
boost::asio::ip::tcp::resolver r(ios);
boost::asio::ip::tcp::socket sock(ios);
boost::asio::connect(sock,
r.resolve(boost::asio::ip::tcp::resolver::query{host, "http"}));
// Send HTTP request using beast
beast::http::request_v1<beast::http::empty_body> req;
req.method = "GET";
req.url = "/";
req.version = 11;
req.headers.replace("Host", host + ":" + std::to_string(sock.remote_endpoint().port()));
req.headers.replace("User-Agent", "Beast");
beast::http::prepare(req);
beast::http::write(sock, req);
// Receive and print HTTP response using beast
beast::streambuf sb;
beast::http::response_v1<beast::http::streambuf_body> resp;
beast::http::read(sock, sb, resp);
std::cout << resp;
}
Upvotes: 4
Reputation: 104464
libcurl docs have an example page on how to get incremental download callbacks (into a memory buffer) as data streams in from a request:
http://curl.haxx.se/libcurl/c/getinmemory.html
In your case, you would just forward the data buffer on to the client that originally made the request.
Upvotes: 5