qtcurl

Run curl_easy transfers asynchronously just inside your Qt main thread.

Behind the scenes there is curl_multi legally hooked on Qt socket & timer events. Not very much of code needed for that actually. But if you're too lazy to deal with curl_multi in event-based system on your own, you'll probably be happy with these sources.

Except the curl_multi thing now there's just only straightforward wraparound for curl_easy functions with a tiny bit of C++/Qt sugar. More things may be added later.

Usage:

1. Set up linking and includes for libcurl

Be strong and do it yourself. There should be plenty of docs for this.

2. Include qtcurl.pri into your .pro

include (qtcurl/src/qtcurl.pri)

3. Write your code

Get started with including the header:

#include "CurlEasy.h"

Create a CurlEasy object and set it up:

CurlEasy *curl = new CurlEasy;
curl->set(CURLOPT_URL, "https://www.google.com");
curl->set(CURLOPT_FOLLOWLOCATION, long(1)); // Tells libcurl to follow HTTP 3xx redirects

You also may want to override some curl callbacks:

curl->setWriteFunction([](char *data, size_t size)->size_t {
    qDebug() << "Data from google.com: " << QByteArray(data, static_cast<int>(size));
    return size;
});

And set some HTTP headers:

curl->setHttpHeader("User-Agent", "My poor little application");

Oh, the signals are there, of course!

QObject::connect(curl, &CurlEasy::done, [curl](CURLcode result) {
    long httpResponseCode = curl->get<long>(CURLINFO_RESPONSE_CODE);
    QString effectiveUrl = curl->get<const char*>(CURLINFO_EFFECTIVE_URL);

    qDebug() << "Transfer for" << effectiveUrl << "is done with"
             << "HTTP" << httpResponseCode
             << "and CURL code" << result;
});

QObject::connect(curl, SIGNAL(done(CURLcode)),
                 curl, SLOT(deleteLater()));

Now it's time to activate the transfer and pass it to all those event-looping things:

curl->perform();
  
// Do your Qt stuff while the request is processing.

Take these usage notes into account:

done() signal will NOT be emitted when the transfer is aborted by abort() method. aborted() will be emitted instead. This is a mostly convenience thing. In the most cases you don't want to do anything in done() when you've aborted the transfer externally.
By default CurlEasy will run on the event loop of the thread from which perform() was called.
Just as almost any QObject stuff, CurlEasy is not generally thread-safe. But just like other QObjects you can spawn it on any thread and safely use there. Even moveToThread is allowed for CurlEasy while the transfer is not running.
All curl_multi-related stuff will be created and set up on the first CurlEasy::perform() call in the current thread.
If you want it to be initialized earlier, just call CurlMulti::threadInstance() to kick that lazy thing.
All curl_multi-related stuff will be destroyed on thread exit (or QApplication exit, see Qt manual for QThreadStorage). If any related CurlEasy transfer has been running at this moment, it will receive aborted() signal.
(Some further notes)

That's all for now. Dig into the sources for details =)

tarasvb/qtcurl

qtcurl

Usage:

1. Set up linking and includes for libcurl

2. Include qtcurl.pri into your .pro

3. Write your code