C++ cURL
Jul 16, 2015

I needed  a quick way to download files for one of my applications and QHttp/QNetworkRequest just wasn’t cutting it. I could have used WinINEet but that’s just a pain to use and you have to handle every possible error that each API call may produce.

cURL offers a fantastic way to easily download files from a URL. It’s incredibly easy to use; I don’t see why anyone would want to use WinINet.

Here are just a few examples on how to use the cURL API. I will add a guide on how to set it up with VC++ later on.

Example 1: Downloading straight to a file

int http_code = 0;

FILE* fp = 0;
fopen_s(&fp, "index.html", "wb");

if(fp)
{
	// Initialize cURL
	CURL* curl = curl_easy_init();
	
	// Set the file to write out to
	curl_easy_setopt(curl, CURLOPT_WRITEDATA, fp);
	
	// Set the URL to download
	curl_easy_setopt(curl, CURLOPT_URL, "www.example.com");
	
	// Download
	curl_easy_perform(curl);
	
	// Get the HTTP response code
	curl_easy_getinfo(curl, CURLINFO_RESPONSE_CODE, &http_code);
	
	// Clean up
	curl_easy_cleanup(curl);
	curl_global_cleanup();
	fclose(fp);
}

Example 2: Downloading to an std::string

size_t curl_writefunc(void* ptr, size_t size, size_t nmemb, std::string* data)
{
	data->append((const char*)ptr, size * nmemb);
	return size * nmemb;
}

int http_code = 0;
std::string data;

// Initialize cURL
CURL* curl = curl_easy_init();

// Set the function to call when there is new data
curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, curl_writefunc);

// Set the parameter to append the new data to
curl_easy_setopt(curl, CURLOPT_WRITEDATA, &data);

// Set the URL to download
curl_easy_setopt(curl, CURLOPT_URL, "www.example.com");

// Download
curl_easy_perform(curl);

// Get the HTTP response code
curl_easy_getinfo(curl, CURLINFO_RESPONSE_CODE, &http_code);

// Clean up
curl_easy_cleanup(curl);
curl_global_cleanup();

Example 3: Transfer the entire file into a vector

struct xfer
{
	// Receive buffer
	std::vector<uint8_t> data;

	// Constructor
	xfer()
	{
	}

	// Destructor
	~xfer()
	{
	}

	// Call this before starting a new transfer
	void clear()
	{
		data.clear();
	}
};

size_t writefunc(const uint8_t* ptr, size_t size, size_t nmemb, xfer* transfer)
{
	// Actual size of the received data
	size_t real_size = size * nmemb;

	// Copy the new data to the end of the vector
	std::copy(ptr, ptr + real_size, std::back_inserter(transfer->data));

	return real_size;
}

int http_code = 0;
boost::shared_ptr<xfer> transfer(boost::make_shared<xfer>());

// Initialize cURL
CURL* curl = curl_easy_init();

// Set the URL
curl_easy_setopt(curl, CURLOPT_URL, "www.example.com");

// Set the function to call when there is new data
curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, writefunc);

// Set the parameter to append the new data to
curl_easy_setopt(curl, CURLOPT_WRITEDATA, transfer.get());

// Download
curl_easy_perform(curl);

// Get the HTTP response code
curl_easy_getinfo(curl, CURLINFO_RESPONSE_CODE, &http_code);

// Cleanup
curl_easy_cleanup(curl);
curl_global_cleanup();


// Display the contents
printf("%s\n", &transfer->data[0]);

You can also add timeouts to prevent cURL from blocking. The code below will set a 5 second timeout.

curl_easy_setopt(curl, CURLOPT_CONNECTTIMEOUT, 5);
Comments