libproxyfish.c~

Linux系统下采用Gtk 2图形库开发的代理搜索验证软件.采用Glib线程池,多线程并发搜索验证,支持国际化,具备中文英文两种语言.

/***************************************************************************
 *            libproxyfish.c
 *
 *  Sat Aug 25 10:10:25 2007
 *  Copyright  2007  hunbuso
 *  [e-mail]
 ****************************************************************************/

#include "main.h"
#include "gui.h"
#include "libproxyfish.h"

static GMutex*		treeview_window_main_proxy_list_mutex = NULL;		/* 信号量 */

static gchar*	get_content_symbol (gchar* content, const gchar* symbol);
static GList*	find_proxy_in_content (gchar* content);
static void 	proxy_fish_thread_run (gchar* data, gpointer user_data);

/**
 * 查找符号
 */
static gchar*	get_content_symbol (gchar* content, const gchar* symbol)
{
	gchar* invector = NULL;
	
	invector = g_strstr_len(content, strlen(content), symbol);
	
	if (!invector)
		return NULL;
	else 
		return invector;
}

/**
 *	描述：检验字符串是否为纯数字
 *	返回：如果字符串是纯数字，则返回真，否则返回假
 *	参数：
 *		data:	检验的字符串
 */
static gboolean check_digital (gchar*  data)
{
	if ((!data) || (0 == strlen(data)))
		return FALSE;
	
	guint	length = strlen(data);
	gint	i;
	
	gchar	*p = data;
	
	for (i = 0; i < length; i++) {
		if (!g_ascii_isdigit(*p))
			return FALSE;
	}
	
	return TRUE;
}

/**
 *	检索	xxx.xxx.xxx.xxx:xxxx格式
 */
static gboolean	find_ip_port_in_content (gchar* content)
{
	const	gchar*	single_point = ".";
	const	gchar*	colon = ":";
	
	gchar*	invector = NULL;
	gchar*	invector_1 = NULL;
	
	gint	i;
	gchar	*j = NULL;
	
	if (!content)
		return FALSE;
	
	invector = content;
	
	///找点号
	for (i = 0; i < 3; i++) {
		invector_1 = get_content_symbol(invector, single_point);

		if (!invector_1)
			return FALSE;
		
		///三位数
		if (invector_1 - invector > 3)
			return FALSE;

		///检验纯数
		for (j = invector; j < invector_1; j++) {
//			DBG("%c\n", *j);
			if (!g_ascii_isdigit(*j))
				return FALSE;
		}
	invector = invector_1 + 1;
	}
	
	///找冒号
	invector_1 = get_content_symbol(invector, colon);
	if (!invector_1)
		return FALSE;
	///三位数
	if (invector_1 - invector > 3)
		return FALSE;
	///检验纯数
	for (j = invector; j < invector_1; j++) {
//		DBG("%c\n", *j);
		if (!g_ascii_isdigit(*j))
			return FALSE;
	}
	
	return TRUE;
}

/**
 *	检索出格式合适的代理
 */
static GList*	find_proxy_in_content (gchar* content)
{
	const	gchar*	single_quotes = "'";
	gchar*	invector = NULL;
	gchar*	invector_1 = NULL;
	gchar*	buffer = NULL;
//	GList*	proxy_list = NULL;
	
	if (!content)
		return NULL;
	
	invector = content;
	
	while (invector) {
		
		invector = get_content_symbol(invector, single_quotes);
		if (!invector)
			break;
		
		invector ++;
		
		invector_1 = get_content_symbol(invector, single_quotes);
		if (!invector_1)
			break;
		
		///此处申请的内存在线程池运行函数proxy_fish_thread_run中释放
		buffer = g_malloc0(invector_1 - invector + 1);
		
		memcpy(buffer, invector, invector_1 - invector);
	
		if (find_ip_port_in_content(buffer)) {
			g_assert(app->env.thread_pool != NULL);
			g_thread_pool_push(app->env.thread_pool, buffer, NULL);
			DBG("%s\n", buffer);
		}
		
		invector = invector_1;
	}
	
	return NULL;
}

/**
 *	描述: 检索具备一定前缀名的文件
 *	返回：具备则返回真，否则返回假
 *	参数：
 *		filename:	文件名
 *		prefix:		前缀，用分号隔开， 例如："http://;ftp://;"
 */
static gboolean filter_file_has_prefix	(const gchar* filename, const gchar* prefix)
{
	gchar 	**prefix_array = g_strsplit(prefix, ";", 0);
	gboolean	result = FALSE;
	
	gint	i = 0;

	while (prefix_array[i])
	{
		if (strlen(prefix_array[i]) > 0)
		{
			if (g_str_has_prefix(filename, prefix_array[i])) {
				result = TRUE;
				break;
			}
		}	
		i++;
	}
	
	g_strfreev(prefix_array);
	
	return result;
}

/**
 *	描述： 检索具备一定后缀名的文件
 *	返回： 具备则返回真，否则返回假
 *  参数： 
 *		filename:	文件名
 *		suffix:		后缀，用冒号隔开，例如："php:html:htm"
 */
static gboolean  filter_file_type	(const gchar* filename, const gchar* suffix)
{
	gchar 	**suffix_array = g_strsplit(suffix, ":", 0);
	gboolean	result = FALSE;
	
	gint	i = 0;

	while (suffix_array[i])
	{
		if (strlen(suffix_array[i]) > 0)
		{
			if (g_str_has_suffix(filename, suffix_array[i])) {
				result = TRUE;
				break;
			}
		}	
		i++;
	}
	
	g_strfreev(suffix_array);
	
	return result;
}

/**
 *	获取内容的子链接
 */
void	get_content_child (gchar* content)
{
//	GList*	url_list = NULL;				///结果
	gchar*	url_keyword = "href=";		///超链接关键字
	gchar*	invector = NULL;
	gchar*  invector_1 = NULL;
	gchar*	buffer = NULL;
	gchar*	symbol = "\"";

	invector = content;
	
	while (invector) {
	invector = g_strstr_len(invector, strlen(content), url_keyword);	
		if (!invector)
			break;
	invector = get_content_symbol(invector, symbol) + 1;
	invector_1 = get_content_symbol(invector, symbol);
		if (!invector_1)
			break;
	buffer = g_malloc0(invector_1 - invector + 1);
	memcpy(buffer, invector, invector_1 - invector);
//	url_list = g_list_append(url_list, buffer);
	if (filter_file_type(buffer, "php:html")){
		DBG("%s\n", buffer);
		g_assert(app->env.thread_pool != NULL);
		g_thread_pool_push(app->env.thread_pool, buffer, NULL);
	}
	invector = invector_1;
	}

//	return url_list;
}

/**
 *	获取并分析网页的内容
 */
gchar*	parse_url_content	(const gchar*	url, enum method mt)
{
	gint 		fd;
	gchar*	init = "GET  HTTP/1.1\r\n"
				  "Host: www.proxycn.com\r\n"
				  "User-Agent: ProxyFish\r\n"
				  "Accept: text/xml,application/xml,application/xhtml+xml,text/html;q=0.9,text/plain;q=0.8,image/png,*/*\r\n"
				  "Keep-Alive: 300\r\n"
				  "Connection: keep-alive\r\n"
				  "Proxy-Connection: keep-alive\r\n"
				  "\r\n";
	gchar*	host_name = "www.proxycn.com";
	GString	*request = NULL;
	guint	delay = 6;			///超时
	struct 	sockaddr_in server_addr;
	struct 	hostent 	*host;
		
	///网站返回的缓冲
	gchar		buf[10*BUFFER_8192];
	memset(buf, 0, sizeof(buf));
	
	request = g_string_new(init);
	request = g_string_insert(request, strlen("GET "), url);
	DBG("%s\n", request->str);
	
	fd = socket(AF_INET, SOCK_STREAM, 0);
    if (fd < 0) {
      perror("socket");
      exit(1);
    }
	if (!(host = gethostbyname(host_name))) {
		perror("gethostbyname");
		exit(1);
	}	

	server_addr.sin_family = AF_INET;
	server_addr.sin_addr=*((struct in_addr *)host->h_addr);
	server_addr.sin_port = htons(80);
	
	bzero(&(server_addr.sin_zero), 8);
	
	if (connect(fd, (struct sockaddr *)&server_addr, sizeof(struct sockaddr)) == -1)	
	{
		perror("connect error");
		close (fd);
		///线程中会用到此函数，这里使用的是返回而不是退出
		return NULL;
	}
	
	struct timeval timeo_test = {delay,0};
	gint savefl = fcntl(fd,F_GETFL);
	fcntl(fd, F_SETFL, savefl);
	setsockopt(fd, SOL_SOCKET, SO_RCVTIMEO, &timeo_test, sizeof(timeo_test));
	
	if (send(fd, request->str, strlen(request->str), 0) < 0) {
		perror("send");
		close (fd);
		///线程中会用到此函数，这里使用的是返回而不是退出
		return NULL;
	}

	while (recv(fd, buf, sizeof(buf),0) > 0) {
		if (get_child_url == mt) {	
//			printf("%s\n",buf);
			get_content_child(buf);
		}
		else if (get_proxy == mt) {
//			printf("%s\n",buf);
			find_proxy_in_content(buf);
		}
	}

	close (fd);
	
	g_string_free(request, TRUE);
	
	return NULL;
}

/*
 *	检验远程服务器的返回值是否含关键字,含有返回真
 */
gboolean	proxy_fish_check_key_string (const char*	buffer, const char* keystr)
{