wip script output verification working

+ trim secure pos segfault
+ concat script message with server message

Makefile

@@ -5,7 +5,7 @@ CXX			 =	c++
 CXXFLAGS	 =	-Wall -Wextra #-Werror
 CXXFLAGS	+=	$(HEADERS_D:%=-I%)
 CXXFLAGS	+=	-std=c++98
-CXXFLAGS	+=	-g
+CXXFLAGS	+=	-g3
 CXXFLAGS	+=	-MMD -MP #header dependencie
 #CXXFLAGS	+=	-O3
 
@@ -29,6 +29,7 @@ SRCS		 =	main.cpp \
 				utils.cpp \
 				cgi_script.cpp \
 				Client.cpp \
+				parsing_message_http.cpp \
 
 OBJS_D		 =	builds
 OBJS		 =	$(SRCS:%.cpp=$(OBJS_D)/%.o)

README.md

@@ -1,23 +1,33 @@
 
 ## work together
 
+#### message
+- err access line response.cpp:169
+- g pas fini la gestion de sortie du cgi (comparaison des headers fields pour supprimer les doublons)
+- g rendu independantes les fonctions qui parses un message http (requete ou response) dans `parsing_http_message.cpp`
+- il y a une erreur importante dans la fonction qui parse les headers d'un message http : ca ne verifie pas si la premier ligne est un header ou la "first line"
+
+#### TODO
+- `_is_cgi()` and `_fill_cgi_path()`
+- `_cgi_output()` change status in `client->status`
+- two cgi tests :
+?	- a basic form with "name" and "something", that return a html page with that
+?		- for GET and POST
+?	- a script called by a file extension in URI
+
 #### output cgi script :
 !	TODO : change all the '\n' by '\r\n'
 !	TODO : there is at least one header field followed by '\r\n\r\n' :
 			- "Content-Type"
 			- "Location"
 			- "Status"
-->	TODO : there is no field duplicate (resolve conflicts)
-->	TODO : if status field, change server status for this one
 !	TODO : there is no space between filed name and ":"
-!	TODO : if no Location field && no Status field -> status code = 200
 !	TODO?: handle Location field, either :
 			- local : start with '/' --> rerun the request with new uri
 			- client : start with '<scheme>:' --> send back status code 302
-
-#### what cgi ?
-- a basic form with "name" and "something", that return a html page with that
-- a script called by a file extension in URI
+->	TODO : there is no field duplicate (resolve conflicts)
+->	TODO : if status field, change server status for this one
+->	TODO : if no Location field && no Status field -> status code = 200
 
 #### questions
 - in client.cpp i fill the port, is there a default one in case it's not in the request ?

big.config

@@ -16,6 +16,7 @@ server {
 	location /board {
 		allow_methods GET;
 		root ./www/html;
+		cgi_ext php cgi
 	}
 
 	location /board/content {

srcs/Client.cpp

@@ -36,10 +36,10 @@ Client::~Client() {
 Client::Client( Client const & src )
 	: body_size(0)
 	, status(0)
-	, _fd		( src.get_cl_fd() )
-	, _port		( src.get_cl_port() )
-	, _ip		( src.get_cl_ip() )
-	, _lsocket	( src.get_cl_lsocket() )
+	, _fd		( src._fd )
+	, _port		( src._port )
+	, _ip		( src._ip )
+	, _lsocket	( src._lsocket )
 {
 	raw_request = src.raw_request;
 	response = src.response;
@@ -70,22 +70,15 @@ Client &	Client::operator=( Client const & rhs )
 // https://www.tutorialspoint.com/http/http_requests.htm
 void	Client::parse_request()
 {
-	std::string							sub;
-	std::vector<std::string>			list;
-	size_t								pos;
+	std::map<std::string, std::string>	headers;
+	std::string							body;
 
 // DEBUG
-std::cout	<< "\nREQUEST _____________________\n"
-			<< raw_request
-			<< "\n\n";
+std::cout << "\nREQUEST ____________\n" << raw_request << "\n_____________\n";
 
-	pos = (raw_request).find(CRLF CRLF);
-	sub = (raw_request).substr(0, pos);
-	list = split(sub, '\n');
-	_parse_request_line(*list.begin());
-	list.erase(list.begin());
-	_parse_request_headers(list);
-	_parse_request_body(pos +  4);
+	_parse_request_line();
+	_parse_request_headers();
+	_parse_request_body();
 	_parse_port_hostname(this->get_rq_headers("Host"));
 	raw_request.clear();
 }
@@ -143,10 +136,10 @@ void	Client::clear_script()
  *********************************************/
 
 // client side
-int					 Client::get_cl_fd()		const { return _fd; }
-const std::string	&Client::get_cl_ip() 		const { return _ip; }
-const std::string	&Client::get_cl_port()		const { return _port; }
-const listen_socket *Client::get_cl_lsocket()	const { return _lsocket; }
+int					 Client::get_cl_fd()	const { return _fd; }
+const std::string	&	Client::get_cl_ip() 		const { return _ip; }
+const std::string	&	Client::get_cl_port()		const { return _port; }
+const listen_socket *	Client::get_cl_lsocket()	const { return _lsocket; }
 
 // requette
 http_method 	Client::get_rq_method()		const { return _request.method; }
@@ -176,30 +169,24 @@ std::string		Client::get_rq_headers(const std::string & key) const
  * PRIVATE MEMBER FUNCTIONS
  *********************************************/
 
-void	Client::_parse_request_line( std::string rline )
+void	Client::_parse_request_line()
 {
-	std::vector<std::string>	sline;
-	std::string					tmp;
+	std::vector<std::string>	line;
+	int							ret;
 
-	sline = split(rline, ' ');
-	if (sline.size() != 3)
+	ret = ::parse_http_first_line(raw_request, line);
+	if (ret != 3)
 	{
-		std::cerr << "err _parse_request_line(): ";
-		throw std::runtime_error("bad request-line header");
+		std::cerr << "err _parse_first_line(): wrong number of elements (" << ret << " instead of 3)\n";
+		status = 400; // "bad request"
+	}
+	else
+	{
+		_request.method = str_to_http_method(line[0]);
+		_request.uri = line[1];
+		_parse_request_uri(line[1]);
+		_request.version = line[2];
 	}
-	// method
-	tmp = ::trim(sline[0], ' ');
-	tmp = ::trim(tmp, '\r');
-	_request.method = str_to_http_method(tmp);
-	// uri
-	tmp = ::trim(sline[1], ' ');
-	tmp = ::trim(tmp, '\r');
-	_request.uri = tmp;
-	_parse_request_uri( tmp );
-	// http version
-	tmp = ::trim(sline[2], ' ');
-	tmp = ::trim(tmp, '\r');
-	_request.version = tmp;
 }
 
 void	Client::_parse_request_uri( std::string uri )
@@ -214,32 +201,16 @@ void	Client::_parse_request_uri( std::string uri )
 	_request.abs_path = uri.substr(0, pos);
 }
 
-void	Client::_parse_request_headers( std::vector<std::string> list )
+void	Client::_parse_request_headers()
 {
-	std::string							key;
-	std::string							val;
-	std::vector<std::string>::iterator	it;
-	size_t								pos;
-
-	for (it = list.begin(); it != list.end(); it++)
-	{
-		pos = (*it).find(':');
-		key = (*it).substr( 0, pos );
-		key = ::trim(key, ' ');
-		key = ::trim(key, '\r');
-		key = ::str_tolower(key);
-		val = (*it).substr( pos + 1 );
-		val = ::trim(val, ' ');
-		val = ::trim(val, '\r');
-		_request.headers.insert( std::pair<std::string, std::string>(key, val) );
-	}
+	// TODO: check error and adjust status
+	_request.headers = ::parse_http_headers(raw_request);
 }
 
-void	Client::_parse_request_body( size_t pos )
+void	Client::_parse_request_body()
 {
-	std::string body = &raw_request[pos];
-
-	_request.body = body;
+	// TODO: check error and adjust status
+	_request.body = ::parse_http_body(raw_request);
 }
 
 void	Client::_parse_port_hostname(std::string host)
@@ -247,10 +218,7 @@ void	Client::_parse_port_hostname(std::string host)
 	size_t	pos;
 
 	if (host == "")
-	{
 		std::cerr << "no host\n";
-		throw std::runtime_error("no host in request");
-	}
 
 	pos = host.find(':');
 	// port :

srcs/Client.hpp

@@ -7,9 +7,10 @@
 # include <map>
 # include <vector>
 # include <string.h>		// strdup
-# include "utils.hpp"
 # include <netinet/in.h>	// sockaddr_in, struct in_addr
 # include <arpa/inet.h>		// htonl, htons, ntohl, ntohs, inet_addr, inet_ntoa
+# include "utils.hpp"
+# include "parsing_message_http.hpp"
 
 struct Script
 {
@@ -74,16 +75,16 @@ class Client
 		bool	fill_script_path(std::string script);
 
 	private:
-		const int				_fd;
-		const std::string		_port;
-		const std::string		_ip;
-		const listen_socket *	_lsocket;
-		struct Request			_request;
+		int				_fd;
+		std::string		_port;
+		std::string		_ip;
+		listen_socket *	_lsocket;
+		struct Request	_request;
 
-		void	_parse_request_line( std::string rline );
+		void	_parse_request_line();
+		void	_parse_request_headers();
+		void	_parse_request_body();
 		void	_parse_request_uri( std::string uri );
-		void	_parse_request_headers( std::vector<std::string> list );
-		void	_parse_request_body( size_t pos );
 		void	_parse_port_hostname(std::string host);
 
 };

srcs/cgi-bin/php-cgi

@@ -1,9 +1,9 @@
 #! /usr/bin/php
 
 <?php
+	echo   "Status: 200\r\n";
+	echo "\r\n";
 	echo "BEGIN PHP-CGI\n-----------\n\n";
-	//phpinfo();
-
 	echo   "AUTH_TYPE:         " . getenv("AUTH_TYPE");
 	echo "\nCONTENT_LENGTH:    " . getenv("CONTENT_LENGTH");
 	echo "\nCONTENT_TYPE:      " . getenv("CONTENT_TYPE");

srcs/utils.cpp

@@ -15,8 +15,19 @@ std::vector<std::string> split(std::string input, char delimiter)
 
 std::string	trim(std::string str, char c)
 {
-	str = str.substr(str.find_first_not_of(c));
-	str = str.substr(0, str.find_last_not_of(c) + 1);
+	size_t	pos;
+
+	// delete leadings c
+	pos = str.find_first_not_of(c);
+	if (pos == std::string::npos)
+		return str;
+	str = str.substr(pos);
+
+	// delete endings c
+	pos = str.find_last_not_of(c);
+	if (pos == std::string::npos)
+		return str;
+	str = str.substr(0, pos + 1);
 
 	return str;
 }
@@ -100,15 +111,37 @@ void	replace_all_substr(std::string &str, const std::string &ori_substr, const s
 	}
 }
 
-bool	operator==(const listen_socket& lhs, int fd)
-	{ return lhs.fd == fd; }
-
-bool	operator==(int fd, const listen_socket& rhs)
-	{ return fd == rhs.fd; }
-
 std::string	str_tolower(std::string str)
 {
 	std::transform(str.begin(), str.end(), str.begin(), ::tolower);
 	return str;
 }
 
+void    del_line_in_str(std::string * str, size_t pos, std::string delim)
+{
+	size_t	begin;
+	size_t	end;
+
+	begin = (*str).rfind(delim, pos);
+	if (begin == std::string::npos)
+		begin = 0;
+	else
+		begin += delim.size();
+
+	end = (*str).find(delim, pos);
+	if (end == std::string::npos)
+		end = 0;
+	else
+		end += delim.size();
+
+	(*str).erase(begin, end - begin);
+}
+
+
+
+bool	operator==(const listen_socket& lhs, int fd)
+	{ return lhs.fd == fd; }
+
+bool	operator==(int fd, const listen_socket& rhs)
+	{ return fd == rhs.fd; }
+

srcs/utils.hpp

@@ -49,5 +49,6 @@ http_method					str_to_http_method(std::string &str);
 std::string					http_methods_to_str(unsigned int methods);
 void						replace_all_substr(std::string &str, const std::string &ori_substr, const std::string &new_substr);
 std::string					str_tolower(std::string str);
+void						del_line_in_str(std::string * str, size_t pos, std::string delim);
 
 #endif

srcs/webserv/Webserv.hpp

@@ -32,7 +32,7 @@
 
 extern bool	g_run;
 extern int	g_last_signal;
-void	signal_handler(int signum);
+void		signal_handler(int signum);
 
 // these might only be TMP
 # define FAILURE -1
@@ -93,14 +93,16 @@ class Webserv
 
 		ServerConfig	&_determine_process_server(Client *client);
 		LocationConfig	&_determine_location(ServerConfig &server, std::string const &path);
-		void	_response_correction(Client *client);
 	// cgi_script.cpp
-		bool	_is_cgi(Client *client);
-		void	_exec_cgi(Client *client);
-		char**	_set_env(Client *client);
-		char*	_dup_env(std::string var, std::string val);
-		char*	_dup_env(std::string var, int i);
-		void	_exec_script(Client *client, char **env);
+		bool		_is_cgi(Client *client);
+		std::string	_exec_cgi(Client *client);
+		char**		_set_env(Client *client);
+		char*		_dup_env(std::string var, std::string val);
+		char*		_dup_env(std::string var, int i);
+		std::string	_exec_script(Client *client, char **env);
+		void		_check_script_output(Client *client, std::string output);
+		void		_check_script_status(Client *client, std::string output);
+		void		_check_script_fields(Client *client, std::string output);
 	// epoll_update.cpp
 		int		_epoll_update(int fd, uint32_t events, int op);
 		int		_epoll_update(int fd, uint32_t events, int op, void *ptr);
@@ -115,6 +117,7 @@ class Webserv
 		void	_listen(int socket_fd, unsigned int max_connections);
 		void	_init_http_status_map();
 		void	_init_mime_types_map();
+
 };
 
 #endif

srcs/webserv/cgi_script.cpp

@@ -11,16 +11,19 @@ bool	Webserv::_is_cgi(Client *client)
 	return false;
 }
 
-void	Webserv::_exec_cgi(Client *client)
+std::string	Webserv::_exec_cgi(Client *client)
 {
-	char**	env;
+	char**			env;
+	std::string		script_output;
 
 	env = _set_env(client);
-	_exec_script(client, env);
+	script_output = _exec_script(client, env);
 
 	for (int i = 0; env[i]; i++)
 		delete[] env[i];
 	delete[] env;
+
+	return script_output;
 }
 
 char*	Webserv::_dup_env(std::string var, std::string val = "")
@@ -49,7 +52,7 @@ char**	Webserv::_set_env(Client *client)
 	env[0]  = _dup_env("AUTH_TYPE");		// authentification not supported
 	env[1]  = _dup_env("CONTENT_LENGTH"		, client->get_rq_body().size());
 	env[2]  = _dup_env("CONTENT_TYPE"		, client->get_rq_headers("Content-Type"));
-	env[3]  = _dup_env("GATEWAY_INTERFACE"	, "CGI/1.0");
+	env[3]  = _dup_env("GATEWAY_INTERFACE"	, "CGI/1.1"); // https://www.rfc-editor.org/rfc/rfc3875
 	env[4]  = _dup_env("PATH_INFO"			, client->get_rq_script_info());
 	env[5]  = _dup_env("PATH_TRANSLATED");	// not supported
 	env[6]  = _dup_env("QUERY_STRING"		, client->get_rq_query());
@@ -69,7 +72,7 @@ char**	Webserv::_set_env(Client *client)
 	return env;
 }
 
-void	Webserv::_exec_script(Client *client, char **env)
+std::string	Webserv::_exec_script(Client *client, char **env)
 {
 	#define RD 0
 	#define WR 1
@@ -82,7 +85,7 @@ void	Webserv::_exec_script(Client *client, char **env)
 	pid_t			pid;
 	char			buf[CGI_BUF_SIZE]; // WIP define buffer
 	char * const *	nll = NULL;
-	std::string		response;
+	std::string		script_output;
 	std::string		body = client->get_rq_body();
 	int				fd_in[2];
 	int				fd_out[2];
@@ -99,7 +102,11 @@ void	Webserv::_exec_script(Client *client, char **env)
 		close(FD_RD_FR_CHLD);
 		dup2(FD_RD_FR_PRNT, STDIN_FILENO);
 		dup2(FD_WR_TO_PRNT, STDOUT_FILENO);
+		// DEBUG
+		std::cerr << "execve:\n";
 		execve(client->get_rq_script_path().c_str(), nll, env);
+		// for tests execve crash :
+		//execve("wrong", nll, env);
 		std::cerr << "execve crashed.\n";
 	}
 	else
@@ -113,19 +120,59 @@ void	Webserv::_exec_script(Client *client, char **env)
 		memset(buf, '\0', CGI_BUF_SIZE);
 		while (read(FD_RD_FR_CHLD, buf, CGI_BUF_SIZE - 1) > 0)
 		{
-			response += buf;
+			script_output += buf;
 			memset(buf, '\0', CGI_BUF_SIZE);
 		}
+		close(FD_RD_FR_CHLD);
+	}
+	if (script_output.empty())
+		script_output = "Status: 500\r\n\r\n";
+	return script_output;
+}
+
+void	Webserv::_check_script_output(Client *client, std::string output)
+{
+	// TODO: it doesn't work with execve error, i don't know why yet ?
+	_check_script_status(client, output);
+	_check_script_fields(client, output);
+}
+
+void	Webserv::_check_script_status(Client *client, std::string output)
+{
+	size_t	pos;
+	int		status_pos;
+
+	pos = output.find("Status:");
+	if (pos != std::string::npos)
+	{
+		status_pos = pos + std::string("Status:").size();
+		client->status = atoi(output.c_str() + status_pos);
+		::del_line_in_str(&output, pos, CRLF);
+	}
+	client->status = 200;
+}
+
+void	Webserv::_check_script_fields(Client *client, std::string output)
+{
+	std::map<std::string, std::string>				srv_fld;	// server_field
+	std::map<std::string, std::string>				scr_fld;	// script_field
+	std::map<std::string, std::string>::iterator	it_srv;
+	std::map<std::string, std::string>::iterator	it_scr;
+	size_t											pos;
+
+	srv_fld = parse_http_headers(client->response);
+	scr_fld = parse_http_headers(output);
+	// wip: compare both map to supress duplicates
+	for (it_srv = srv_fld.begin(); it_srv != srv_fld.end(); it_srv++)
+	{
+		for (it_scr = scr_fld.begin(); it_scr != scr_fld.end(); it_scr++)
+		{
+			if (it_srv->first == it_scr->first)
+			{
+				pos = client->response.find(it_srv->first);
+				::del_line_in_str(&client->response, pos, CRLF);
+			}
+		}
 	}
-	if (response.empty())
-		response = "Status: 500\r\n\r\n";
-
-// DEBUG
-std::cout	<< "\n_______response_______\n"
-			<< response
-			<< "\n_____end response_____\n";
-
-	// TODO: see how this must be handled
-	client->response += response;
 }
 

srcs/webserv/parsing_message_http.cpp

@@ -0,0 +1,90 @@
+
+#include "parsing_message_http.hpp"
+
+size_t
+	parse_http_first_line(std::string message, std::vector<std::string> &line)
+{
+	std::vector<std::string>	sline;
+	std::string					sub;
+	std::string					tmp;
+	size_t						pos;
+	size_t						ret;
+
+	// TODO: check for err in substr
+	pos = message.find(CRLF);
+	sub = message.substr(0, pos);
+	sline = ::split(sub, ' ');
+	ret = sline.size();
+	if (ret != 3)
+		return ret;
+	for (int i = 0; i < 3; i++)
+	{
+		tmp = ::trim(sline[i], ' ');
+		tmp = ::trim(tmp, '\r');
+		line.push_back(tmp);
+	}
+	return ret;
+}
+
+std::map<std::string, std::string>
+	parse_http_headers(std::string message)
+{
+	std::map<std::string, std::string>	headers;
+	std::vector<std::string>			list;
+	std::vector<std::string>::iterator	it;
+	std::string							sub;
+	std::string							key;
+	std::string							val;
+	size_t								pos;
+
+	pos = (message).find(CRLF CRLF);
+	sub = (message).substr(0, pos);
+	list = ::split(sub, '\n');
+	if ( maybe_http_first_line( *list.begin() ) )
+		list.erase(list.begin());
+
+	for (it = list.begin(); it != list.end(); it++)
+	{
+		// TODO: if pattern is not "NAME: value" return error
+		pos = (*it).find(':');
+		key = (*it).substr( 0, pos );
+		key = ::trim(key, ' ');
+		key = ::trim(key, '\r');
+		key = ::str_tolower(key);
+		val = (*it).substr( pos + 1 );
+		val = ::trim(val, ' ');
+		val = ::trim(val, '\r');
+		headers.insert( std::pair<std::string, std::string>(key, val) );
+	}
+	return headers;
+}
+
+std::string
+	parse_http_body(std::string message)
+{
+	std::string							body;
+	size_t								pos;
+
+	pos = message.find(CRLF CRLF);
+	pos += std::string(CRLF CRLF).size();
+	// TODO: copying just like that might fail in case of binary or images
+	body = message.substr(pos);
+
+	return body;
+}
+
+bool	maybe_http_first_line(std::string str)
+{
+//	method  SP target SP version	https://www.rfc-editor.org/rfc/rfc7230.html#section-3.1.1
+//	version SP status SP reason		https://www.rfc-editor.org/rfc/rfc7230.html#section-3.1.2
+
+	std::vector<std::string>	sline;
+
+	sline = ::split(str, ' ');
+	if (sline.size() != 3)
+		return false;
+	if (sline[0].find(':') != std::string::npos)
+		return false;
+	return true;
+}
+

srcs/webserv/parsing_message_http.hpp

@@ -0,0 +1,36 @@
+
+#ifndef PARSING_MESSAGE_HTTP_HPP
+# define PARSING_MESSAGE_HTTP_HPP
+
+# include <iostream>
+# include <string>
+# include <vector>
+# include <map>
+# include "utils.hpp"
+
+size_t
+	parse_http_first_line(std::string message, std::vector<std::string> &line);
+
+std::map<std::string, std::string>
+	parse_http_headers(std::string message);
+
+std::string
+	parse_http_body(std::string message);
+
+bool
+	maybe_http_first_line(std::string);
+
+//	http message structure :
+//
+//	start-line
+//		request-line
+//			method SP target SP version
+//		response-line
+//			version SP status SP reason
+//	header-fields
+//		name ":" SP value
+//	CRLF
+//	body
+
+#endif
+

srcs/webserv/parsing_request.cpp

@@ -0,0 +1,2 @@
+#include "parsing_request.hpp"
+

srcs/webserv/response.cpp

@@ -145,10 +145,15 @@ void	Webserv::_get(Client *client, ServerConfig &server, LocationConfig &locatio
 
 	// TMP HUGO
 	//
+	std::string	script_output;
 	if (_is_cgi(client))
 	{
-		_exec_cgi(client);
-		_response_correction(client);
+		script_output = _exec_cgi(client);
+		// DEBUG
+		std::cout << "\n____script_output____\n" << script_output << "\n_______________\n";
+		// wip check output of script
+		_check_script_output(client, script_output);
+		client->response += script_output;
 		return;
 	}
 	//
@@ -224,12 +229,6 @@ void	Webserv::_get_file(Client *client, const std::string &path)
 	}
 }
 
-// WIP HUGO
-void	Webserv::_response_correction(Client *client)
-{
-	(void)client;
-}
-
 void	Webserv::_append_body(Client *client, const char *body, size_t body_size, const std::string &file_extension)
 {
 /*

www/form_get.html

@@ -0,0 +1,13 @@
+<!DOCTYPE html>
+<html>
+	<head>
+		<meta charset="UTF-8">
+		<title></title>
+	</head>
+	<body>
+		<form method="get">
+			<input type="submit" value="submit">
+		</form>
+	</body>
+</html>
+

www/form_post.html

@@ -0,0 +1,13 @@
+<!DOCTYPE html>
+<html>
+	<head>
+		<meta charset="UTF-8">
+		<title></title>
+	</head>
+	<body>
+		<form method="post">
+			<input type="submit" value="submit">
+		</form>
+	</body>
+</html>
+