premiere-libtorrent/include/libtorrent/http_parser.hpp

182 lines
6.1 KiB
C++
Raw Normal View History

2008-01-12 10:36:03 +01:00
/*
2014-02-23 20:12:25 +01:00
Copyright (c) 2008-2014, Arvid Norberg
2008-01-12 10:36:03 +01:00
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
* Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in
the documentation and/or other materials provided with the distribution.
* Neither the name of the author nor the names of its
contributors may be used to endorse or promote products derived
from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef TORRENT_HTTP_PARSER_HPP_INCLUDED
#define TORRENT_HTTP_PARSER_HPP_INCLUDED
#include <map>
#include <string>
#include <utility>
#include <vector>
2008-01-12 10:36:03 +01:00
#ifdef _MSC_VER
#pragma warning(push, 1)
#endif
#include <boost/cstdint.hpp>
#include <boost/tuple/tuple.hpp>
#ifdef _MSC_VER
#pragma warning(pop)
#endif
#include "libtorrent/config.hpp"
#include "libtorrent/buffer.hpp"
#include "libtorrent/size_type.hpp"
namespace libtorrent
{
2010-09-21 08:34:13 +02:00
// return true if the status code is 200, 206, or in the 300-400 range
bool is_ok_status(int http_status);
// return true if the status code is a redirect
bool is_redirect(int http_status);
2014-09-01 10:59:12 +02:00
TORRENT_EXTRA_EXPORT std::string resolve_redirect_location(std::string referrer
, std::string location);
class TORRENT_EXTRA_EXPORT http_parser
2008-01-12 10:36:03 +01:00
{
public:
enum flags_t { dont_parse_chunks = 1 };
http_parser(int flags = 0);
2012-03-19 07:06:52 +01:00
~http_parser();
2008-01-12 10:36:03 +01:00
std::string const& header(char const* key) const
{
static std::string empty;
std::multimap<std::string, std::string>::const_iterator i
2008-01-12 10:36:03 +01:00
= m_header.find(key);
if (i == m_header.end()) return empty;
return i->second;
}
std::string const& protocol() const { return m_protocol; }
int status_code() const { return m_status_code; }
std::string const& method() const { return m_method; }
std::string const& path() const { return m_path; }
std::string const& message() const { return m_server_message; }
buffer::const_interval get_body() const;
bool header_finished() const { return m_state == read_body; }
bool finished() const { return m_finished; }
boost::tuple<int, int> incoming(buffer::const_interval recv_buffer
, bool& error);
int body_start() const { return m_body_start_pos; }
size_type content_length() const { return m_content_length; }
std::pair<size_type, size_type> content_range() const
{ return std::make_pair(m_range_start, m_range_end); }
2008-01-12 10:36:03 +01:00
// returns true if this response is using chunked encoding.
// in this case the body is split up into chunks. You need
// to call parse_chunk_header() for each chunk, starting with
// the start of the body.
bool chunked_encoding() const { return m_chunked_encoding; }
// removes the chunk headers from the supplied buffer. The buffer
// must be the stream received from the http server this parser
// instanced parsed. It will use the internal chunk list to determine
// where the chunks are in the buffer. It returns the new length of
// the buffer
int collapse_chunk_headers(char* buffer, int size) const;
// returns false if the buffer doesn't contain a complete
// chunk header. In this case, call the function again with
// a bigger buffer once more bytes have been received.
// chunk_size is filled in with the number of bytes in the
// chunk that follows. 0 means the response terminated. In
// this case there might be additional headers in the parser
// object.
// header_size is filled in with the number of bytes the header
// itself was. Skip this number of bytes to get to the actual
// chunk data.
// if the function returns false, the chunk size and header
// size may still have been modified, but their values are
// undefined
bool parse_chunk_header(buffer::const_interval buf
, size_type* chunk_size, int* header_size);
// reset the whole state and start over
2008-01-12 10:36:03 +01:00
void reset();
bool connection_close() const { return m_connection_close; }
std::multimap<std::string, std::string> const& headers() const { return m_header; }
std::vector<std::pair<size_type, size_type> > const& chunks() const { return m_chunked_ranges; }
2008-01-12 10:36:03 +01:00
private:
size_type m_recv_pos;
2008-01-12 10:36:03 +01:00
std::string m_method;
std::string m_path;
std::string m_protocol;
std::string m_server_message;
size_type m_content_length;
size_type m_range_start;
size_type m_range_end;
2008-01-12 10:36:03 +01:00
std::multimap<std::string, std::string> m_header;
2008-01-12 10:36:03 +01:00
buffer::const_interval m_recv_buffer;
// contains offsets of the first and one-past-end of
// each chunked range in the response
std::vector<std::pair<size_type, size_type> > m_chunked_ranges;
// while reading a chunk, this is the offset where the
// current chunk will end (it refers to the first character
// in the chunk tail header or the next chunk header)
size_type m_cur_chunk_end;
2014-07-06 21:18:00 +02:00
int m_status_code;
// the sum of all chunk headers read so far
int m_chunk_header_size;
int m_partial_chunk_header;
// controls some behaviors of the parser
int m_flags;
2014-07-06 21:18:00 +02:00
int m_body_start_pos;
enum { read_status, read_header, read_body, error_state } m_state;
// this is true if the server is HTTP/1.0 or
// if it sent "connection: close"
bool m_connection_close;
bool m_chunked_encoding;
bool m_finished;
2008-01-12 10:36:03 +01:00
};
}
#endif // TORRENT_HTTP_PARSER_HPP_INCLUDED