Parse HTTP headers in C++ - c++

Question

I am using curl to communicate with a server.
When I make a request for data I receive the HTTP headers followed by jpeg data separated by a boundary like so:
I need to parse out
The boundary string
The Content-Length.
I have copied the incoming data to a a char array like so:
static size_t OnReceiveData ( void * pvData, size_t tSize, size_t tCount, void * pvUser )
{
printf("%*.*s", tSize * tCount, tSize * tCount, pvData);
char* _data;
if(pvData != nullptr && 0 != tCount)
{
_data = new char[tCount];
memcpy(_data, pvData, tCount);
}
return ( tCount );
}
How can I best do this in C++?? How do I actually inspect and parse the _data array for the information that I want?? Are the any boost libraries that I can use for example??

Answer 1

You could parse the headers on the fly or put them into a map and post-process later.
Use find, substr methods from the std::string.
Look at Boost String Algorithms Library, it contains lots of algorithms, e.g. trim
e.g. to place headers into the std::map and print them (rough cuts):
#include <stdlib.h>
#include <iostream>
#include <sstream>
#include <string>
#include <map>
#include <boost/algorithm/string.hpp>
int main(int argc, char* argv[]) {
const char* s = "HTTP/1.1 200 OK\r\n"
"Content-Type: image/jpeg; charset=utf-8\r\n"
"Content-Length: 19912\r\n\r\n";
std::map<std::string, std::string> m;
std::istringstream resp(s);
std::string header;
std::string::size_type index;
while (std::getline(resp, header) && header != "\r") {
index = header.find(':', 0);
if(index != std::string::npos) {
m.insert(std::make_pair(
boost::algorithm::trim_copy(header.substr(0, index)),
boost::algorithm::trim_copy(header.substr(index + 1))
));
}
}
for(auto& kv: m) {
std::cout << "KEY: `" << kv.first << "`, VALUE: `" << kv.second << '`' << std::endl;
}
return EXIT_SUCCESS;
}
You will get the output:
KEY: `Content-Length`, VALUE: `19912`
KEY: `Content-Type`, VALUE: `image/jpeg; charset=utf-8`
Having the headers, you could extract the required ones for post-processing.

Answer 2

I would put all headers in a map, after which you can easily iterate through it. No boost needed. Here a basic working example with libcurl:
#include <iostream>
#include <string>
#include <map>
#include <curl/curl.h>
static size_t OnReceiveData (void * pData, size_t tSize, size_t tCount, void * pmUser)
{
size_t length = tSize * tCount, index = 0;
while (index < length)
{
unsigned char *temp = (unsigned char *)pData + index;
if ((temp[0] == '\r') || (temp[0] == '\n'))
break;
index++;
}
std::string str((unsigned char*)pData, (unsigned char*)pData + index);
std::map<std::string, std::string>* pmHeader = (std::map<std::string, std::string>*)pmUser;
size_t pos = str.find(": ");
if (pos != std::string::npos)
pmHeader->insert(std::pair<std::string, std::string> (str.substr(0, pos), str.substr(pos + 2)));
return (tCount);
}
int main(int argc, char* argv[])
{
CURL *curl = curl_easy_init();
if (!curl)
return 1;
std::map<std::string, std::string> mHeader;
curl_easy_setopt(curl, CURLOPT_URL, "http://www.example.com");
curl_easy_setopt(curl, CURLOPT_HEADERFUNCTION, OnReceiveData);
curl_easy_setopt(curl, CURLOPT_HEADERDATA, &mHeader);
curl_easy_setopt(curl, CURLOPT_NOBODY, true);
curl_easy_perform(curl);
curl_easy_cleanup(curl);
std::map<std::string, std::string>::const_iterator itt;
for (itt = mHeader.begin(); itt != mHeader.end(); itt++)
{
if (itt->first == "Content-Type" || itt->first == "Content-Length")
std::cout << itt->first << ": " << itt->second << std::endl;
}
}

Answer 3

The cpp-netlib project (based on boost) contains a full MIME parser (written with boost.spirit).
I'm not really that happy with the interface of the parser, but it works well.