Resume broken https download in libcurl - c++

Question

I have a binary file to be downloaded and using curl in linux terminal, the following command supports download resume on a broken request.
curl -C - -o sample1.bin https://speed.hetzner.de/100MB.bin
The above will resume a download that is canceled.
When i use libcurl in my cpp program to do the same, is there any api's that i can use to achieve the above result on a HTTPS broken request.
Thank you for you help.
NOTE: CURL_RESUME_FROM does not have support on HTTPS.

Answer 1

I think you could implement a retry system by yourself, for example:
CURL *curl;
curl = curl_easy_init();
//Set curl options as needed with curl_easy_setopt()
char* url;
int tries = 0;
bool done = false;
while (tries != 3 && !done) {
res = curl_easy_perform(curl);
curl_easy_getinfo(curl, CURLINFO_RESPONSE_CODE, &http_code);
curl_easy_getinfo(curl, CURLINFO_TOTAL_TIME, &elapsed);
curl_easy_getinfo(curl, CURLINFO_EFFECTIVE_URL, &url);
if (res != CURLE_OK || http_code != 200)
tries++;
else
done = true;
}
// Check if any error has occurred
if (res != CURLE_OK || http_code != 200) {
// Could not perform request "
if (tries == 3) {
//Too many tries, remote host is overloaded or down
} else {
// Cannot perform CURL
}
}
// Curl succeeded
Also you can have a look to CURLOPT_LOW_SPEED_LIMIT and CURLOPT_LOW_SPEED_TIME to avoid any overhead in the server.

Answer 2

How about something "low-level" like passing an extra header? Let s be the size of the partially-downloaded file, just use Range: bytes=s-.
See Requesting_a_specific_range_from_a_server and CURLOPT_HTTPHEADER explained.
#include <curl/curl.h>
#include <string>
#include <sstream>
int dim=... //size of partial download sample1.bin
std::string s=std::to_string(dim); // <-- here s is the string representing the size of the partial download
CURL *curl = curl_easy_init();
struct curl_slist *list = NULL;
if(curl) {
curl_easy_setopt(curl, CURLOPT_URL,"https://speed.hetzner.de/100MB.bin");
list = curl_slist_append(list, "Range: bytes="+s+"-"); //from where it left off to the end (or where it stops again)
curl_easy_setopt(curl, CURLOPT_HTTPHEADER, list);
curl_easy_perform(curl);
curl_slist_free_all(list); /* free the list again */
}
Check out also Making HTTPS GET with libcurl.