Multithreaded curl application has memory allocation problems - c++

I'm working on an application in C++ that threads and hands a bunch of threads URLs for cURL to download in parallel.
I'm employing a method that should be safe to download images and videos, etc. I uses memcpy instead of assuming the data is a string or character array.
I pass each thread a structure, thread_status, for a number of things. The structure lets the parent process know the thread is done downloading. It also stores the data cURL is downloading and keeps track of it's size as cURL returns more buffers for writing.
I pass a (void *) pointer that points to each structure that's allocated at initialization to each thread that does the downloading. The first page is downloaded properly, after that I keep getting errors from realloc().
Here is the simplest example that illustrates my problem. This sample is not multi-threaded but uses a similar structure to keep track of itself.
#include <string>
#include <assert.h>
#include <iostream>
#include <curl/curl.h>
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#define NOT_READY 1
#define READY 0
using namespace std;
struct thread_status {
int id;
pthread_t *pid;
int readyState;
char *url;
void *data;
size_t bufferlen;
size_t writepos;
int initialized;
} ;
size_t static
writefunction( void *ptr, size_t size,
size_t nmemb, void *userdata)
{
size_t nbytes = size*nmemb;
struct thread_status **this_status;
this_status = (struct thread_status **) userdata;
if (!(*this_status)->initialized){
(*this_status)->data = (void *)malloc(1024);
(*this_status)->bufferlen = 1024;
(*this_status)->writepos = 0;
(*this_status)->initialized = true;
}
if ((*this_status)->bufferlen < ((*this_status)->writepos + nbytes)){
(*this_status)->bufferlen = (*this_status)->bufferlen + nbytes;
(*this_status)->data = realloc((*this_status)->data, (size_t) ((*this_status)->writepos + nbytes));
}
assert((*this_status)->data != NULL);
memcpy((*this_status)->data + (*this_status)->writepos, ptr, nbytes);
(*this_status)->writepos += nbytes;
return nbytes;
}
void *pull_data (void *my_struct){
struct thread_status *this_struct;
this_struct = (struct thread_status *) my_struct;
this_struct->initialized = false;
cout<<(char *)this_struct->url<<"\n";
CURL *curl;
curl = curl_easy_init();
size_t rc = 0;
while(true){
curl_easy_setopt(curl,
CURLOPT_WRITEFUNCTION, writefunction);
curl_easy_setopt(curl,
CURLOPT_WRITEDATA, (void *) &this_struct);
curl_easy_setopt(curl,
CURLOPT_NOSIGNAL, true);
curl_easy_setopt(curl,
CURLOPT_URL, (char *)this_struct->url);
if (curl_easy_perform(curl) != 0){
cout<<"curl did not perform\n";
exit(1);
} else {
if (this_struct->data != NULL){
// Use a binary write.
rc = fwrite(this_struct->data, this_struct->writepos, 1, stdout);
free(this_struct->data);
} else {
cout<<"Data is NULL\n";
}
}
// Tell the babysitter the thread is ready.
this_struct->readyState = READY;
// This would pause the thread until the parent thread has processed the data in it.
// while(this_struct->readyState == READY){;}
// Now get ready for another round!
this_struct->writepos = (size_t) 0;
this_struct->initialized = false;
this_struct->bufferlen = (size_t) 0;
break;
}
curl_easy_cleanup(curl);
return (void *)"a";
}
int main(){
char *urls[] = { "http://www.example.com/", "http://www.google.com", "http://www.touspassagers.com/", "http://www.facebook.com/" };
int i=0;
struct thread_status mystatuses[4];
for (i=0;i<4;i++){
struct thread_status my_status;
char *data;
my_status.id = i;
my_status.readyState = NOT_READY;
my_status.url = urls[i];
my_status.data = data;
my_status.bufferlen = 0;
my_status.writepos = 0;
my_status.initialized = false;
mystatuses[i] = my_status;
}
for (i=0;i<4;i++){
cout<<"pulling #"<<i<<"\n";
pull_data((void *)&mystatuses[i]);
}
}
If anyone can enlighten me as to the source of my error or a remedy for it I would appreciate it.

You might consider using valgrind to help locate the source of the memory problems.

Got it!
Apparently 1KB isn't enough memory to handle the first cURL buffer. I changed 1024 to nbytes and it works!
Before the memory memcpy put in the buffer ran over the allocated memory resulting in corruptions.
I did a post about it if anyone cares to see the full implementation:
http://www.touspassagers.com/2011/01/a-working-curlopt_writefunction-function-for-libcurl/

Related

write access violation with c++ libcurl [closed]

Closed. This question is not reproducible or was caused by typos. It is not currently accepting answers.
This question was caused by a typo or a problem that can no longer be reproduced. While similar questions may be on-topic here, this one was resolved in a way less likely to help future readers.
Closed last month.
Improve this question
I'm trying to create c++ code that download data from some URLs, but it's throwing a write access violation:
#include <errno.h>
#include <stdlib.h>
#include <string.h>
#ifndef WIN32
#endif
#include <curl/curl.h>
#include <string>
static const char* urls[] = {
"http://www.example.com",
"http://www.example1.com",
};
#define MAX_PARALLEL 10 /* number of simultaneous transfers */
#define NUM_URLS sizeof(urls)/sizeof(char *)
static size_t write_cb(void* ptr, size_t size, size_t nmemb, void* buffer)
{
((std::string*)buffer)->append((char*)ptr, nmemb);
return nmemb;
}
static void add_transfer(CURLM* cm, int i, int* left)
{
CURL* eh = curl_easy_init();
curl_easy_setopt(eh, CURLOPT_WRITEFUNCTION, write_cb);
curl_easy_setopt(eh, CURLOPT_URL, urls[i]);
curl_easy_setopt(eh, CURLOPT_PRIVATE, urls[i]);
curl_easy_setopt(eh, CURLOPT_WRITEDATA, &write_cb);
curl_easy_setopt(eh, CURLOPT_VERBOSE, 1L);
curl_multi_add_handle(cm, eh);
(*left)++;
}
int main(void)
{
CURLM* cm;
unsigned int transfers = 0;
int msgs_left = -1;
int left = 0;
curl_global_init(CURL_GLOBAL_ALL);
cm = curl_multi_init();
/* Limit the amount of simultaneous connections curl should allow: */
curl_multi_setopt(cm, CURLMOPT_MAXCONNECTS, (long)MAX_PARALLEL);
for (transfers = 0; transfers < MAX_PARALLEL && transfers < NUM_URLS;
transfers++)
add_transfer(cm, transfers, &left);
do {
int still_alive = 1;
curl_multi_perform(cm, &still_alive);
CURLMsg* msg;
int queued;
CURLMcode mc = curl_multi_perform(cm, &still_alive);
if (cm)
/* wait for activity, timeout or "nothing" */
mc = curl_multi_poll(cm, NULL, 0, 1000, NULL);
if (mc)
break;
do {
msg = curl_multi_info_read(cm, &queued);
if (msg) {
if (msg->msg == CURLMSG_DONE) {
/* a transfer ended */
fprintf(stderr, "Transfer completed\n");
}
}
} while (msg);
if (left)
curl_multi_wait(cm, NULL, 0, 1000, NULL);
} while (left);
curl_multi_cleanup(cm);
curl_global_cleanup();
return EXIT_SUCCESS;
}
Its crashing on the line:
_Mypair._Myval2._Mysize = _Old_size + _Count;
The full error message is:
Exception thrown: write access violation.
this was 0x7FF7941D39D0.
How can I make this code download each Url data without any error?
In your write_cb() callback, you are expecting the buffer parameter to point at a std::string object, but in add_transfer() you are setting CURLOPT_WRITEDATA to point at write_cb itself rather than at a std::string object.
Try something more like this instead:
struct url_info {
const char* url;
std::string data;
};
static url_info urls[] = {
{"http://www.example.com", ""},
{"http://www.example1.com", ""}
};
static const int NUM_URLS = sizeof(urls)/sizeof(urls[0]);
static size_t write_cb(void* ptr, size_t size, size_t nmemb, void* buffer)
{
size_t result = size * nmemb;
static_cast<std::string*>(buffer)->append(static_cast<char*>(ptr), result);
return result;
}
static void add_transfer(CURLM* cm, int i, int* left)
{
...
curl_easy_setopt(eh, CURLOPT_URL, urls[i].url);
curl_easy_setopt(eh, CURLOPT_WRITEDATA, &urls[i].data);
curl_easy_setopt(eh, CURLOPT_PRIVATE, &urls[i]);
...
}
int main(void)
{
...
while ((msg = curl_multi_info_read(cm, &queued)) != NULL) {
if (msg->msg == CURLMSG_DONE) {
/* a transfer ended */
url_info *info;
curl_easy_getinfo(msg->easy_handle, CURLINFO_PRIVATE, reinterpret_cast<char**>(&info));
// use info->url and info->data as needed...
std::cerr << "Transfer completed from " << info->url << ", bytes received: " << info->data.size() << "\n";
}
}
...
}

json conversion error c++ IsString failed

I have data in a JSON object that I can't seem to get to.
Error:
> testGetprice3.o: testGetprice3.cpp:71: int getData(): Assertion
> `document["success"].IsString()' failed.
I have tried switching the datatypes, it only errors.
How do I correctly access my JSON object data?
#include "rapidjson/include/rapidjson/document.h"
#include <iostream>
#include <fstream>
#include <stdio.h>
#include <curl/curl.h>
#include <unistd.h>
#include <unordered_map>
#include <string>
using namespace rapidjson;
struct myData
{
std::fstream *file;
std::string *str;
};
size_t write_data(void *ptr, size_t size, size_t nmemb, myData *data)
{
size_t numBytes = size * nmemb;
if (data->file)
data->file->write((char*)ptr, numBytes);
if (data->str)
*data->str += std::string((char*)ptr, numBytes);
return numBytes;
}
//function to get coin data and perform analysis
int getData()
{
int count = 0;
//begin non terminating loop
while(true)
{
count++;
CURL *curl = curl_easy_init();
if (curl)
{
curl_easy_setopt(curl, CURLOPT_URL, "http://pubapi.cryptsy.com/api.php?method=singlemarketdata&marketid=155");
std::fstream file("/home/coinz/cryptsy/myfile.txt", std::ios_base::out | std::ios_base::ate);
std::string json;
myData data;
data.file = &file;
data.str = &json;
curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, &write_data);
curl_easy_setopt(curl, CURLOPT_WRITEDATA, &data);
/* Perform the request, res will get the return code */
CURLcode res = curl_easy_perform(curl);
/* Check for errors */
if (res != CURLE_OK)
{
std::cerr << "curl_easy_perform() failed: " << curl_easy_strerror(res) << std::endl;
}
else
{
file << std::endl;
//begin deserialization
Document document;
document.Parse(json.c_str());
assert(document.HasMember("success"));
assert(document["success"].IsString());
//std::cout << "The Last Traded Price is = " << document["lasttradeprice"].GetString() << std::endl;
}
/* always cleanup */
curl_easy_cleanup(curl);
}
//timer for URL request. *ADUJST ME AS DESIRED*
usleep(10000000);
}
return 0;
}
//Le Main
int main(void)
{
getData();
}
Looking at your JSON data, the success member is an int (as it is not surrounded by quote marks).
More info on JSON types can be found here: http://en.wikipedia.org/wiki/JSON#Data_types.2C_syntax_and_example
If the success member was a string, the JSON data would look like this: {"success":"1",...
Therefore when you call assert(document["success"].IsString()) it causes the error, as the success member is not a string, but an int.
Since the success member is an int, change the assertion to test whether it's an int - assert(document["success"].IsInt())
You stated you changed datatypes, but did not state which ones you tried. From the URL you are using, success is being returned as "success":1, which is an int.
Change:
assert(document["success"].IsString());
to:
assert(document["success"].IsInt());

Creating a C++ class to read data from a website to a string using libcurl

I am using http://ubuntuforums.org/showthread.php?t=781021 as a guide for how to write a c++ program that use libcurl to check a website and downloads the text to a c++ app. When I copy the code to my visual studio project, it works with only slight modification. However, I hit a snag trying to convert it to class based files.
I created a .h and .cpp files to run the query. What I didn't realize before was that handle_data was not a typical function, since it is not given any arguments or even () in the midway down Viewer.cpp. This is how the above site did it, and it worked, but I'm at a loss as to why or how to convert it now.
Does anyone know whats going on here or how to fix it?
Thanks!
Viewer.h:
#pragma once
#include <string>
#include "curl.h"
class Viewer
{
public:
Viewer(void);
~Viewer(void);
std::string view(std::string q);
private:
size_t handle_data(void *ptr, size_t size, size_t nmemb, void *stream);
std::string contents;
};
Viewer.cpp:
#include "stdafx.h"
#include "Viewer.h"
#include <iostream>
Viewer::Viewer(void)
{
std::cout << "ViewerCreated!\n";
}
Viewer::~Viewer(void)
{
}
size_t Viewer::handle_data(void *ptr, size_t size, size_t nmemb, void *stream)
{
int numbytes = size*nmemb;
// The data is not null-terminated, so get the last character, and replace
// it with '\0'.
char lastchar = *((char *) ptr + numbytes - 1);
*((char *) ptr + numbytes - 1) = '\0';
contents.append((char *)ptr);
contents.append(1,lastchar);
*((char *) ptr + numbytes - 1) = lastchar; // Might not be necessary.
return size*nmemb;
}
std::string Viewer::view(std::string q)
{
char* url = "www.google.com";
CURL* curl = curl_easy_init();
if(curl)
{
// Tell libcurl the URL
curl_easy_setopt(curl,CURLOPT_URL, url);
// Tell libcurl what function to call when it has data
curl_easy_setopt(curl,CURLOPT_WRITEFUNCTION,handle_data);
// Do it!
CURLcode res = curl_easy_perform(curl);
curl_easy_cleanup(curl);
if (res == 0)
std::cout << contents << std::endl;
else
std::cerr << "Error: " << res << std::endl;
}
return contents;
}
You should not replace the last character of the chunk that you have received. Instead you should force that the buffer containing the appended blocks has a '\0' just after the end of the data. This way it can be printed, but if you are going to store that in a file, you should just write the bytes that you have received.
This has always worked for me:
struct MemoryStruct chunk;
chunk.memory=NULL;
chunk.size = 0;
...
curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, WriteMemoryCallback);
curl_easy_setopt(curl, CURLOPT_WRITEDATA, (void *)&chunk);
And the callback:
struct MemoryStruct {
char *memory;
size_t size;
};
static void *myrealloc(void *ptr, size_t size)
{
if(ptr)
return realloc(ptr, size);
else
return malloc(size);
}
static size_t WriteMemoryCallback(void *ptr, size_t size, size_t nmemb, void *data)
{
size_t realsize = size * nmemb;
struct MemoryStruct *mem = (struct MemoryStruct *)data;
mem->memory = (char *)myrealloc(mem->memory, mem->size + realsize + 1);
if (mem->memory) {
memcpy(&(mem->memory[mem->size]), ptr, realsize);
mem->size += realsize;
mem->memory[mem->size] = 0;
}
return realsize;
}
handle_data is used as a parameter for curl_easy_setopt. What that means is that you have to program what handle_data does when it receives the next data chunk (providing the data chunk is what curl will do for you).
handle_data is being passed to curl_easy_setopt as a callback function, which it will call when there's data available. Because of that, handle_data must be called as a static function (that is, one not receiving an object pointer, because cURL doesn't have one). However, you're not declaring it as static. Even if you do, you are accessing contents inside it, which is invalid.

Adding to char array isn't working

I'm trying to read a text file line by line, and add each line to a char array. But the lines aren't added, at all.
//This is the default char array that comes with the cURL code.
char *text[]={
"one\n",
"two\n",
"three\n",
" Hello, this is CURL email SMTP\n",
NULL
};
/*Now we're going to replace that char array, with an array that holds the contents of a textfile.
We'll read a textfile out line by line, and add each line to the char array.
*/
void makemailmessage()
{
text[0] = '\0'; //Clear text
text[0] = "testy\n"; //First line in new char array
//Read the text file, add each line to the char array.
string line;
ifstream myfile ("C:\\Users\\admin\\Downloads\\bbb.txt");
int counter;
counter = 1;
if (myfile.is_open())
{
while ( myfile.good() )
{
getline (myfile,line);
//Convert the string variable "line" to a char (a)
char *a=new char[line.size()+1];
a[line.size()]=0;
memcpy(a,line.c_str(),line.size());
//Add \n to the end of "a" (new char will be "str")
char str[80];
strcpy (str,a);
strcat (str,"\n");
//Add "str" to the char array "text"
text[counter] = str;
text[counter+1] = "test\n"; //Also added this for testing purposes
write_data("C:\\Users\\admin\\Downloads\\checkit.txt", str); //Also for testing purposes
//Increase counter by 2 because we added two new items to the char array "text"
counter++;
counter++;
}
myfile.close();
text[counter-1] = "testy2\n"; //Ad another text line
text[counter] = NULL; //End char array
}
Each str is written correctly to checkit.txt but for some reason it is not added to the char array because I end up with the char array looking like this:
testy
test
test
testy2
What am I doing wrong?
UPDATE2:
The reason I am trying to make a char array is because the cURL function I am using needs a char array to form the email body. This is the important part of the cURL code.
static size_t read_callback(void *ptr, size_t size, size_t nmemb, void *userp)
{
struct WriteThis *pooh = (struct WriteThis *)userp;
const char *data;
if(size*nmemb < 1)
return 0;
data = text[pooh->counter]; //This part is using the char array.
if(data) {
size_t len = strlen(data);
memcpy(ptr, data, len);
pooh->counter++;
return len;
}
return 0;
}
Here's the full code
Okay, after chatting on this a bit more, here is a fix:
C++ version
Full code file here: https://gist.github.com/1342118#file_test.cpp
Replace the relevant code with:
#include <vector>
#include <fstream>
// ...
std::vector<std::string> text;
static int read_text(char* fname)
{
//Read the text file, add each line to the char array.
std::ifstream myfile (fname);
std::string line;
while (std::getline(myfile, line))
text.push_back(line + '\n');
return 0;
}
static size_t read_callback(void *ptr, size_t size, size_t nmemb, void *userp)
{
/* This was already in. */
struct WriteThis *pooh = (struct WriteThis *)userp;
if(size*nmemb < 1)
return 0;
if (pooh->counter < text.size())
{
const std::string& data = text[pooh->counter];
memcpy(ptr, data.data(), data.length());
pooh->counter++; /* advance pointer */
return data.length();
}
return 0; /* no more data left to deliver */
}
Pure C version
Full code file here: https://gist.github.com/1342118#file_test.c
Replace
//This is the default char array that comes with the cURL code.
char *text[]={
"one\n",
"two\n",
"three\n",
" Hello, this is CURL email SMTP\n",
NULL
};
With
char **text = 0;
static int read_text(char* fname)
{
unsigned capacity = 10;
int linecount = 0;
// free_text(); see below
text = realloc(text, capacity*sizeof(*text));
FILE* file = fopen(fname, "r");
if (!file)
{ perror("Opening file"); return 1; }
char buf[2048];
char* line = 0;
while (line = fgets(buf, sizeof(buf), file))
{
if (linecount>=capacity)
{
capacity *= 2;
text = realloc(text, capacity*sizeof(*text));
}
text[linecount++] = strdup(line);
}
fclose(file);
return 0;
}
Hook it up in you main function, e.g. like so
if (argc<2)
{
printf("Usage: %s <email.eml>\n", argv[0]);
exit(255);
} else
{
printf("Reading email body from %s\n", argv[1]);
if (0 != read_text(argv[1]))
exit(254);
}
Or, if you so prefer, just calling read_text("C:\\Users\\admin\\Downloads\\bbb.txt") :)
To really top things off, don't forget to reclaim memory when you're done - properly:
#include "curl/curl.h"
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <memory.h>
#include <string.h>
#define GetCurrentDir getcwd
#define USERNAME "obscured#gmail.com"
#define PASSWORD "obscured"
#define SMTPSERVER "smtp.gmail.com"
#define SMTPPORT ":587"
#define RECIPIENT "<obscured#gmail.com>"
#define MAILFROM "<obscured#gmail.com>"
#define MULTI_PERFORM_HANG_TIMEOUT 60 * 1000
/* Note that you should include the actual meta data headers here as well if
you want the mail to have a Subject, another From:, show a To: or whatever
you think your mail should feature! */
char **text = 0;
void free_text()
{
if (text)
{
char** it;
for (it = text; *it; ++it)
free(*it);
free(text);
text = 0;
}
}
static int read_text(char* fname)
{
unsigned capacity = 10;
int linecount = 0;
free_text();
text = realloc(text, capacity*sizeof(*text));
FILE* file = fopen(fname, "r");
if (!file)
{ perror("Opening file"); return 1; }
char buf[2048];
char* line = 0;
while (line = fgets(buf, sizeof(buf), file))
{
if (linecount>=capacity)
{
capacity *= 2;
text = realloc(text, capacity*sizeof(*text));
}
text[linecount++] = strdup(line);
}
if (linecount>=capacity)
text = realloc(text, (++capacity)*sizeof(*text));
text[linecount] = 0; // terminate
fclose(file);
return 0;
}
struct WriteThis {
int counter;
};
static size_t read_callback(void *ptr, size_t size, size_t nmemb, void *userp)
{
/* This was already in. */
struct WriteThis *pooh = (struct WriteThis *)userp;
const char *data;
if(size*nmemb < 1)
return 0;
data = text[pooh->counter];
if(data) {
size_t len = strlen(data);
memcpy(ptr, data, len);
pooh->counter++; /* advance pointer */
return len;
}
return 0; /* no more data left to deliver */
}
static struct timeval tvnow(void)
{
/*
** time() returns the value of time in seconds since the Epoch.
*/
struct timeval now;
now.tv_sec = (long)time(NULL);
now.tv_usec = 0;
return now;
}
static long tvdiff(struct timeval newer, struct timeval older)
{
return (newer.tv_sec-older.tv_sec)*1000+
(newer.tv_usec-older.tv_usec)/1000;
}
int main(int argc, char** argv)
{
if (argc<2)
{
printf("Usage: %s <email.eml>\n", argv[0]);
exit(255);
} else
{
printf("Reading email body from %s\n", argv[1]);
if (0 != read_text(argv[1]))
exit(254);
}
CURL *curl;
CURLM *mcurl;
int still_running = 1;
struct timeval mp_start;
char mp_timedout = 0;
struct WriteThis pooh;
struct curl_slist* rcpt_list = NULL;
pooh.counter = 0;
curl_global_init(CURL_GLOBAL_DEFAULT);
curl = curl_easy_init();
if(!curl)
return 1;
mcurl = curl_multi_init();
if(!mcurl)
return 2;
rcpt_list = curl_slist_append(rcpt_list, RECIPIENT);
/* more addresses can be added here
rcpt_list = curl_slist_append(rcpt_list, "<others#example.com>");
*/
curl_easy_setopt(curl, CURLOPT_URL, "smtp://" SMTPSERVER SMTPPORT);
curl_easy_setopt(curl, CURLOPT_USERNAME, USERNAME);
curl_easy_setopt(curl, CURLOPT_PASSWORD, PASSWORD);
curl_easy_setopt(curl, CURLOPT_READFUNCTION, read_callback);
curl_easy_setopt(curl, CURLOPT_MAIL_FROM, MAILFROM);
curl_easy_setopt(curl, CURLOPT_MAIL_RCPT, rcpt_list);
curl_easy_setopt(curl, CURLOPT_USE_SSL, CURLUSESSL_ALL);
curl_easy_setopt(curl, CURLOPT_SSL_VERIFYPEER,0);
curl_easy_setopt(curl, CURLOPT_SSL_VERIFYHOST, 0);
curl_easy_setopt(curl, CURLOPT_READDATA, &pooh);
curl_easy_setopt(curl, CURLOPT_VERBOSE, 1);
curl_easy_setopt(curl, CURLOPT_SSLVERSION, 0);
curl_easy_setopt(curl, CURLOPT_SSL_SESSIONID_CACHE, 0);
curl_multi_add_handle(mcurl, curl);
mp_timedout = 0;
mp_start = tvnow();
/* we start some action by calling perform right away */
curl_multi_perform(mcurl, &still_running);
while(still_running) {
struct timeval timeout;
int rc; /* select() return code */
fd_set fdread;
fd_set fdwrite;
fd_set fdexcep;
int maxfd = -1;
long curl_timeo = -1;
FD_ZERO(&fdread);
FD_ZERO(&fdwrite);
FD_ZERO(&fdexcep);
/* set a suitable timeout to play around with */
timeout.tv_sec = 1;
timeout.tv_usec = 0;
curl_multi_timeout(mcurl, &curl_timeo);
if(curl_timeo >= 0) {
timeout.tv_sec = curl_timeo / 1000;
if(timeout.tv_sec > 1)
timeout.tv_sec = 1;
else
timeout.tv_usec = (curl_timeo % 1000) * 1000;
}
/* get file descriptors from the transfers */
curl_multi_fdset(mcurl, &fdread, &fdwrite, &fdexcep, &maxfd);
/* In a real-world program you OF COURSE check the return code of the
function calls. On success, the value of maxfd is guaranteed to be
greater or equal than -1. We call select(maxfd + 1, ...), specially in
case of (maxfd == -1), we call select(0, ...), which is basically equal
to sleep. */
//rc = select(maxfd+1, &fdread, &fdwrite, &fdexcep, &timeout);
if (tvdiff(tvnow(), mp_start) > MULTI_PERFORM_HANG_TIMEOUT) {
fprintf(stderr, "ABORTING TEST, since it seems "
"that it would have run forever.\n");
break;
}
switch(rc) {
case -1:
/* select error */
break;
case 0: /* timeout */
default: /* action */
curl_multi_perform(mcurl, &still_running);
break;
}
}
curl_slist_free_all(rcpt_list);
curl_multi_remove_handle(mcurl, curl);
curl_multi_cleanup(mcurl);
curl_easy_cleanup(curl);
curl_global_cleanup();
free_text();
return 0;
}
I'm trying to read a text file line by line, and add each line to a
char array.
Since this is C++, why not use an std::vector<string> and use the std::string version of getline?
The std::string class will look after the memory needed to hold a string of any sort of length, and the std::vector class will worry about the memory needed to hold an "array", so to speak, of strings.
EDIT: Actually looking at your code again, you do use an std::string and then allocate memory to store it as an array of chars, and then store pointers to those strings in some fixed sized array, test. Why go to all that trouble when, as I mentioned above, you can use an std::vector<string> to hold all your std::string objects? Mind = boggled.
EDIT2: Couldn't you also use cURLpp as a C++ wrapper for cURL? I haven't used either so I can't comment on the effectiveness of it.
What am I doing wrong?
For one, this:
char str[80];
strcpy (str,a);
strcat (str,"\n");
//Add "str" to the char array "text"
text[counter] = str;
str is allocated on the stack, with block-wide scope. Then you enter that pointer in an array with a greater scope. This is usually a recipe for disaster - a rather impressive segmentation fault or whatever the equivalent is on your platform.
In this case, due to its use in a loop, your program will either crash, or - if the stars have the proper alignment - you will end up with all pointers in your array pointing to the same out-of-scope string, namely the one that was last read.
Why do you even go into that trouble, when you have already dynamically allocated a in the heap?
By the way, mixing char[] arrays (and the associated standard C library functions) with C++ strings is NOT a good idea. Not even an acceptable one. OK, it a bad idea. Just stick to C++ strings...

Can one read a remote file as an istream with libcurl?

I'd like to use the libcurl library to open a remote date file and iterate through it with an istream. I've looked through the nice example in this thread but it writes the remote file to a local file. Instead I'd like to have the remote reads be pushed to an istream for subsequent programmatic manipulation. Is this possible? I would greatly appreciate help.
Best,
Aaron
Boost's IO Stream might be a better solution than STL's own stream. At least it is much simpler to create a boost stream. From boost's own docs:
#include <curl/curl.h>
#include <boost/iostreams/stream.hpp>
class CURLDevice
{
private:
CURL* handle;
public:
typedef char char_type;
typedef boost::iostreams::source_tag category;
CURLDevice()
{
handle = curl_easy_init();
}
CURLDevice(const std::string &url)
{
handle = curl_easy_init();
open( url );
}
~CURLDevice()
{
curl_easy_cleanup(handle);
}
void open(const std::string &url)
{
curl_easy_setopt(handle, CURLOPT_URL, url.c_str());
curl_easy_setopt(handle, CURLOPT_CONNECT_ONLY, 1);
curl_easy_perform(handle);
}
std::streamsize read(char* s, std::streamsize n)
{
size_t read;
CURLcode ret = curl_easy_recv(handle, s, n, &read);
if ( ret == CURLE_OK || ret == CURLE_AGAIN )
return read;
else
return -1;
}
};
typedef boost::iostreams::stream<CURLDevice> CURLStream;
int main(int argc, char **argv)
{
curl_global_init(CURL_GLOBAL_ALL);
{
CURLStream stream("http://google.com");
char buffer[256];
int sz;
do
{
sz = 256;
stream.read( buffer, sz );
sz = stream.gcount();
std::cout.write( buffer, sz );
}
while( sz > 0 );
}
curl_global_cleanup();
return 0;
}
Note: when I run the code above I get a segfault in CURL, this appears to be because I don't know exactly how to use curl itself.