Question

I’m working on a program which will download lyrics from sites like AZLyrics. I’m using libcurl.

It’s my code

lyricsDownloader.cpp

#include "lyricsDownloader.h"
#include <curl/curl.h>
#include <cstring>
#include <iostream>

#define DEBUG 1

/////////////////////////////////////////////////////////////////////////////


size_t lyricsDownloader::write_data_to_var(char *ptr, size_t size, size_t nmemb, void *userdata) // this function is a static member function
{
    ostringstream * stream = (ostringstream*) userdata;
    size_t count = size * nmemb;
    stream->write(ptr, count);
    return count;
}


string AZLyricsDownloader::toProviderCode() const
{ /*this creates an url*/ }

CURLcode AZLyricsDownloader::download()
{
    CURL * handle;
    CURLcode err;
    ostringstream buff;
    handle = curl_easy_init();
    if (! handle) return static_cast<CURLcode>(-1);
    // set verbose if debug on
    curl_easy_setopt( handle, CURLOPT_VERBOSE, DEBUG );
    curl_easy_setopt( handle, CURLOPT_URL, toProviderCode().c_str() ); // set the download url to the generated one
    curl_easy_setopt(handle, CURLOPT_WRITEDATA, &buff);
    curl_easy_setopt(handle, CURLOPT_WRITEFUNCTION, &AZLyricsDownloader::write_data_to_var);
    err = curl_easy_perform(handle); // The segfault should be somewhere here - after calling the function but before it ends
    cerr << "cleanup\n";
    curl_easy_cleanup(handle);

    // copy the contents to text variable
    lyrics = buff.str();
    return err;
}

main.cpp

#include <QString>
#include <QTextEdit>
#include <iostream>
#include "lyricsDownloader.h"

int main(int argc, char *argv[])
{
        AZLyricsDownloader dl(argv[1], argv[2]);
        dl.perform();
        QTextEdit qtexted(QString::fromStdString(dl.lyrics));
        cout << qPrintable(qtexted.toPlainText());
        return 0;
}

When running

./maelyrica Anthrax Madhouse

I’m getting this logged from curl

* About to connect() to azlyrics.com port 80 (#0)
*   Trying 174.142.163.250... * connected
* Connected to azlyrics.com (174.142.163.250) port 80 (#0)
> GET /lyrics/anthrax/madhouse.html HTTP/1.1
Host: azlyrics.com
Accept: */*

< HTTP/1.1 301 Moved Permanently
< Server: nginx/1.0.12
< Date: Thu, 05 Jul 2012 16:59:21 GMT
< Content-Type: text/html
< Content-Length: 185
< Connection: keep-alive
< Location: http://www.azlyrics.com/lyrics/anthrax/madhouse.html
< 
Segmentation fault

Strangely, the file is there. The same error is displayed when there’s no such page (redirect to azlyrics.com mainpage)

What am I doing wrong?

Thanks in advance

EDIT: I made the function for writing data static, but this changes nothing.
Even wget seems to have problems

$ wget http://www.azlyrics.com/lyrics/anthrax/madhouse.html
--2012-07-06 10:36:05--  http://www.azlyrics.com/lyrics/anthrax/madhouse.html
Resolving www.azlyrics.com... 174.142.163.250
Connecting to www.azlyrics.com|174.142.163.250|:80... connected.
HTTP request sent, awaiting response... No data received.
Retrying.

Why does opening the page in a browser work and wget/curl not?

EDIT2: After adding this:

curl_easy_setopt(handle, CURLOPT_FOLLOWLOCATION, 1);

And making the function static everything’s OK.

Answer 1

Your code

    curl_easy_setopt(handle,CURLOPT_WRITEFUNCTION,&AZLyricsDownloader::write_data_to_var);

and the following quote from the documentation from libcurl

There’s basically only one thing to keep in mind when using C++
instead of C when interfacing libcurl:

The callbacks CANNOT be non-static class member functions

Example C++ code:

class AClass { static size_t write_data(void *ptr, size_t size, size_t nmemb, void* ourpointer) { /* do what you want with the data */ } }

could be the source of your problem as your function is not a static member. Even if not you are breaking this rule.

This may not solve your problem but given the amount of code you have posted in your example, that was the first thing that immediately came to mind and it is worth changing this as recommended by libcurl. If it does not solve your problem I would suggest identifying the error you are getting in more detail so that you can pose a more specific question next time (with a lot less code displayed).