path: root/plugins/fcgi/fcgi.cpp

#include "fcgi.h"

#include "fastcgi.h"

#include <boost/algorithm/string/predicate.hpp>
#include <boost/array.hpp>
#include <boost/asio.hpp>
#include <boost/endian/conversion.hpp>
#include <boost/coroutine2/coroutine.hpp>
#include <boost/process.hpp>

#include <algorithm>
#include <filesystem>
#include <fstream>
#include <iostream>
#include <string>
#include <unordered_map>

using namespace std::string_literals;
namespace bp = boost::process;
namespace fs = std::filesystem;

using boost::asio::ip::tcp;

struct FCGIContext
{
  std::function<std::string(const std::string& key)>& GetServerParam;
  std::function<std::string(const std::string& key)>& GetRequestParam; // request including body (POST...)
  std::function<void(const std::string& key, const std::string& value)>& SetResponseHeader; // to be added to result string

  FCGIContext(std::function<std::string(const std::string& key)>& p_GetServerParam,
             std::function<std::string(const std::string& key)>& p_GetRequestParam,
             std::function<void(const std::string& key, const std::string& value)>& p_SetResponseHeader
             )
   : GetServerParam(p_GetServerParam)
   , GetRequestParam(p_GetRequestParam)
   , SetResponseHeader(p_SetResponseHeader)
 {
 }
};

namespace {

 const std::string gateway_interface{"CGI/1.1"};

 // Return a reasonable mime type based on the extension of a file.
 std::string mime_type(fs::path path)
 {
  using boost::algorithm::iequals;
  auto const ext = [&path]
  {
   size_t pos = path.string().rfind(".");
   if (pos == std::string::npos)
    return std::string{};
   return path.string().substr(pos);
  }();
  if(iequals(ext, ".htm"))  return "text/html"; // TODO: unordered_map
  if(iequals(ext, ".html")) return "text/html";
  if(iequals(ext, ".php"))  return "text/html";
  if(iequals(ext, ".css"))  return "text/css";
  if(iequals(ext, ".txt"))  return "text/plain";
  if(iequals(ext, ".js"))   return "application/javascript";
  if(iequals(ext, ".json")) return "application/json";
  if(iequals(ext, ".xml"))  return "application/xml";
  if(iequals(ext, ".swf"))  return "application/x-shockwave-flash";
  if(iequals(ext, ".flv"))  return "video/x-flv";
  if(iequals(ext, ".png"))  return "image/png";
  if(iequals(ext, ".jpe"))  return "image/jpeg";
  if(iequals(ext, ".jpeg")) return "image/jpeg";
  if(iequals(ext, ".jpg"))  return "image/jpeg";
  if(iequals(ext, ".gif"))  return "image/gif";
  if(iequals(ext, ".bmp"))  return "image/bmp";
  if(iequals(ext, ".ico"))  return "image/vnd.microsoft.icon";
  if(iequals(ext, ".tiff")) return "image/tiff";
  if(iequals(ext, ".tif"))  return "image/tiff";
  if(iequals(ext, ".svg"))  return "image/svg+xml";
  if(iequals(ext, ".svgz")) return "image/svg+xml";
  return "application/text";
 }

 typedef boost::coroutines2::coroutine<std::string> coro_t;

 // returns true iff std::string is empty or contains newline
 bool isEmpty(const std::string& s)
 {
  return s.empty() || s == "\r" || s == "\n"s || s == "\r\n"s;
 }

 void trimLinebreak(std::string& s)
 {
  size_t pos = s.find_last_not_of("\r\n");
  if (pos != s.npos)
   s = s.substr(0, pos + 1);
 }

 std::unordered_map<std::string, std::function<void(std::string&, FCGIContext&)>> headerMap {
  { "CACHE-CONTROL", [](std::string& v, FCGIContext& c){ c.SetResponseHeader("cache_control", v); } },

  { "CONTENT-TYPE", [](std::string& v, FCGIContext& c){ c.SetResponseHeader("content_type", v); } },

  { "SET-COOKIE", [](std::string& v, FCGIContext& c){ c.SetResponseHeader("set_cookie", v); } },
  
  { "STATUS", [](std::string& v, FCGIContext& c) {
      std::string status{"500"};
      if (v.size() >= 3) {
       status = v.substr(0, 3);
      }
      c.SetResponseHeader("status", status);
  } }
 };

 void handleHeader(const std::string& s, FCGIContext& context)
 {
  size_t pos = s.find(": ");
  if (pos == s.npos)
   return;

  std::string key {s.substr(0, pos)};
  std::string value {s.substr(pos + 2)};

  std::transform(key.begin(), key.end(), key.begin(), ::toupper);
  auto it {headerMap.find(key)};
  if (it == headerMap.end())
   std::cout << "Warning: Unhandled CGI header: " << s << std::endl;
  else
   it->second(value, context);
 }

 void setFCGIEnvironment(std::unordered_map<std::string, std::string>& env, FCGIContext& c)
 {
  std::string authorization {c.GetRequestParam("authorization")};
  if (!authorization.empty())
   env["AUTH_TYPE"] = c.GetRequestParam("authorization");

  env["CONTENT_LENGTH"] = c.GetRequestParam("content_length");
  env["CONTENT_TYPE"] = c.GetRequestParam("content_type");
  env["GATEWAY_INTERFACE"] = gateway_interface;

  std::string target {c.GetRequestParam("target")};
  size_t query_pos {target.find("?")};
  std::string query;
  if (query_pos != target.npos) {
   query = target.substr(query_pos + 1);
   target = target.substr(0, query_pos);
  }

  env["PATH_INFO"] = c.GetRequestParam("rel_target");
  env["PATH_TRANSLATED"] = fs::path{c.GetRequestParam("rel_target")} / c.GetRequestParam("rel_target");
  env["QUERY_STRING"] = query;
  env["REMOTE_ADDR"] = "";
  env["REMOTE_HOST"] = "";
  env["REMOTE_IDENT"] = "";
  env["REMOTE_USER"] = "";
  env["REQUEST_METHOD"] = c.GetRequestParam("method");
  env["REQUEST_URI"] = target;
  env["SCRIPT_NAME"] = c.GetRequestParam("target");
  env["SERVER_NAME"] = c.GetRequestParam("host");
  env["SERVER_PORT"] = c.GetServerParam("port");
  env["SERVER_PROTOCOL"] = c.GetRequestParam("http_version");
  env["SERVER_SOFTWARE"] = c.GetServerParam("version");

  env["HTTP_ACCEPT"] = c.GetRequestParam("http_accept");
  env["HTTP_ACCEPT_CHARSET"] = c.GetRequestParam("http_accept_charset");
  env["HTTP_ACCEPT_ENCODING"] = c.GetRequestParam("http_accept_encoding");
  env["HTTP_ACCEPT_LANGUAGE"] = c.GetRequestParam("http_accept_language");
  env["HTTP_CONNECTION"] = c.GetRequestParam("http_connection");
  env["HTTP_HOST"] = c.GetRequestParam("http_host");
  env["HTTP_USER_AGENT"] = c.GetRequestParam("http_user_agent");
  env["HTTP_REFERER"] = c.GetRequestParam("referer");
  env["HTTP_COOKIE"] = c.GetRequestParam("cookie");
  env["HTTPS"] = c.GetRequestParam("https");
 }

 class FCGI_Record
 {
  std::vector<char> m_data;
 public:
  // create record to send
  FCGI_Record(unsigned char type, uint16_t id, unsigned char arg1, unsigned char arg2)
  {
   if (type == FCGI_BEGIN_REQUEST) {
    size_t size {sizeof(FCGI_BeginRequestRecord)};
    m_data.resize(size);
    FCGI_BeginRequestRecord& r{*reinterpret_cast<FCGI_BeginRequestRecord*>(m_data.data())};
    r.header.version = FCGI_VERSION_1;
    r.header.type = type;
    r.header.requestIdB1 = id >> 8;
    r.header.requestIdB0 = id & 0xFF;
    r.header.contentLengthB1 = 0;
    r.header.contentLengthB0 = sizeof(r.body);
    r.body.roleB1 = 0;
    r.body.roleB0 = arg1;
    r.body.flags = arg2;
   } else
    throw std::runtime_error("Bad FCGI type: "s + std::to_string(type));
  }
  
  // create record to send
  FCGI_Record(unsigned char type, uint16_t id, const std::string& data)
  {
   if (type == FCGI_PARAMS || type == FCGI_STDIN) {
    size_t size {sizeof(FCGI_Header) + data.size()};
    m_data.resize(size);
    FCGI_Header& r{*reinterpret_cast<FCGI_Header*>(m_data.data())};
    r.version = FCGI_VERSION_1;
    r.type = type;
    r.requestIdB1 = id >> 8;
    r.requestIdB0 = id & 0xFF;
    r.contentLengthB1 = 0;
    r.contentLengthB0 = data.size();
    memcpy((void*)&m_data[sizeof(FCGI_Header)], (void*)data.data(), data.size());
   } else
    throw std::runtime_error("Bad FCGI type: "s + std::to_string(type));
  }

  // parse record
  FCGI_Record(std::vector<char>& v)
  {
   if (v.size() < sizeof(FCGI_Header))
    throw std::length_error("No full FCGI header available");

   FCGI_Header& r{*reinterpret_cast<FCGI_Header*>(v.data())};
   
   size_t content_length {((size_t)r.contentLengthB1) << 8 | r.contentLengthB0};
   size_t record_size {sizeof(FCGI_Header) + content_length};
   if (v.size() < record_size)
    throw std::length_error("No full FCGI record available");

   m_data = std::vector(v.begin(), v.begin() + record_size);

   v.erase(v.begin(), v.begin() + record_size);
  }

  std::vector<char>& getBuffer()
  {
   return m_data;
  }

  unsigned char getType() { return reinterpret_cast<FCGI_Header*>(m_data.data())->type; }

  std::string getContent()
  {
   if (m_data.size() < sizeof(FCGI_Header))
    throw std::runtime_error("No data available in FCGI_Record: "s + std::to_string(m_data.size()) + " of "s + std::to_string(sizeof(FCGI_Header)) + " bytes"s);
   return std::string(m_data.data() + sizeof(FCGI_Header), m_data.size() - sizeof(FCGI_Header));
  }
 };

 std::string encode_u8(size_t v)
 {
  unsigned char c {static_cast<unsigned char>(v)};
  return std::string(reinterpret_cast<char*>(&c), 1);
 }

 std::string encode_u32(size_t v)
 {
  uint32_t x {static_cast<uint32_t>(v)};
  boost::endian::native_to_big_inplace(x);
  return std::string(reinterpret_cast<char*>(&x), sizeof(x));
 }

 void FCGI_EncodeEnv(const std::unordered_map<std::string, std::string>& map, std::string& s)
 {
  s.clear();

  for (auto&[key, value]: map) {
   if (key.size() > 127)
    s += encode_u32(key.size());
   else
    s += encode_u8(key.size());
   
   if (value.size() > 127)
    s += encode_u32(value.size());
   else
    s += encode_u8(value.size());

   s += key;
   s += value;
  }
 }

 // Used to return errors by generating response page and HTTP status code
 std::string HttpStatus(std::string status, std::string message, std::function<plugin_interface_setter_type>& SetResponseHeader)
 {
  SetResponseHeader("status", status);
  SetResponseHeader("content_type", "text/html");
  return status + " " + message;
 }

} // anonymous namespace

std::string fcgi_plugin::fcgiQuery(FCGIContext& context)
{
 // host:port or unix domain socket
 std::string app_addr{context.GetRequestParam("doc_root")};

 std::string output_data;

 std::unordered_map<std::string, std::string> env;
 setFCGIEnvironment(env, context);
 std::string env_bytes;
 FCGI_EncodeEnv(env, env_bytes);

 size_t pos { app_addr.find(':') };
 if (pos != app_addr.npos) { // host:port
  boost::asio::io_context io_context; // TODO: member?
  tcp::resolver resolver(io_context);
  auto endpoints{resolver.resolve(app_addr.substr(0, pos), app_addr.substr(pos + 1))};
  tcp::socket socket(io_context);
  boost::asio::connect(socket, endpoints);

  if (!socket.is_open()) {
   return HttpStatus("500", "FCGI connection", context.SetResponseHeader);
  }

  FCGI_ID_Guard id_guard(m_fcgi_id);
  uint16_t id{id_guard.getID()};

  FCGI_Record begin_request{FCGI_BEGIN_REQUEST, id, FCGI_RESPONDER, FCGI_KEEP_CONN};
  socket.write_some(boost::asio::buffer(begin_request.getBuffer()));

  FCGI_Record params{FCGI_PARAMS, id, env_bytes};
  socket.write_some(boost::asio::buffer(params.getBuffer()));
 
  FCGI_Record params_end{FCGI_PARAMS, id, std::string{}};
  socket.write_some(boost::asio::buffer(params_end.getBuffer()));

  FCGI_Record stdin_{FCGI_PARAMS, id, context.GetRequestParam("body")};
  socket.write_some(boost::asio::buffer(stdin_.getBuffer()));
  
  FCGI_Record stdin_end{FCGI_PARAMS, id, std::string{}};
  socket.write_some(boost::asio::buffer(stdin_end.getBuffer()));

  bool ended{false};
  std::vector<char> inbuf;
  while (!ended) {
   std::vector<char> inbuf_part(1024);
   size_t got {socket.read_some(boost::asio::buffer(inbuf_part))};
   inbuf.insert(inbuf.end(), inbuf_part.begin(), inbuf_part.begin() + got);

   try {
    FCGI_Record r{inbuf};
    if (r.getType() == FCGI_END_REQUEST) {
     ended = true;
    } else if (r.getType() == FCGI_STDOUT) {
     output_data += r.getContent();
    } else if (r.getType() == FCGI_STDERR) {
     std::cerr << "FCGI Error: " << r.getContent();
    } else
     throw std::runtime_error("Unhandled FCGI type: "s + std::to_string(r.getType()));
   } catch (const std::length_error& ex) {
    // ignore if not enough data available yet
   }
  }
 } else { // Unix domain socket, or file to start
  // TODO
 }

 std::istringstream is_out{output_data};
 std::string output;
 std::string line;

 // TODO: C++20 coroutine 
 coro_t::push_type processLine( [&](coro_t::pull_type& in){
                               std::string line;
                               // read header lines
                               while (in && !isEmpty(line = in.get())) {
                                trimLinebreak(line);
                                handleHeader(line, context);
                                in();
                               }
                               
                               // read empty line
                               if (!isEmpty(line))
                                throw std::runtime_error("Missing empty line between CGI header and body");
                               if (in)
                                in();

                               // read remainder
                               while (in) {
                                line = in.get();
                                output += line + '\n';
                                in();
                               }

                               throw std::runtime_error("Input missing on processing CGI body");
 });

 while (std::getline(is_out, line)) {
  processLine(line);
 }

 return output;
}

std::string fcgi_plugin::name()
{
 return "fcgi";
}

fcgi_plugin::fcgi_plugin()
{
 //std::cout << "Plugin constructor" << std::endl;
}

fcgi_plugin::~fcgi_plugin()
{
 //std::cout << "Plugin destructor" << std::endl;
}

std::string fcgi_plugin::generate_page(
  std::function<std::string(const std::string& key)>& GetServerParam,
  std::function<std::string(const std::string& key)>& GetRequestParam, // request including body (POST...)
  std::function<void(const std::string& key, const std::string& value)>& SetResponseHeader // to be added to result string
)
{
 try {
  // Request path must not contain "..".
  std::string rel_target{GetRequestParam("rel_target")};
  size_t query_pos{rel_target.find("?")};
  if (query_pos != rel_target.npos)
   rel_target = rel_target.substr(0, query_pos);

  std::string target{GetRequestParam("target")};
  if (rel_target.find("..") != std::string::npos) {
   return HttpStatus("400", "Illegal request: "s + target, SetResponseHeader);
  }

  SetResponseHeader("content_type", "text/html");

  FCGIContext context(GetServerParam, GetRequestParam, SetResponseHeader);

  try {
   return fcgiQuery(context);
  } catch (const std::runtime_error& ex) {
   return HttpStatus("404", "Not found: "s + GetRequestParam("target"), SetResponseHeader);
  } catch (const std::exception& ex) {
   return HttpStatus("500", "Internal Server Error: "s + ex.what(), SetResponseHeader);
  }

 } catch (const std::exception& ex) {
  return HttpStatus("500", "Unknown Error: "s + ex.what(), SetResponseHeader);
 }
}