using System.Collections.Generic;
using System.Text;
using System.Net;
using System.Web;
using System.IO;
using System.Collections;
using System.Text.Regularexpressions;
namespace chinaz
{
class Program
{
staticvoid Main(string[] args)
{
string cookie =null;
using (StreamReader sr =new StreamReader("cookie.txt"))
{
cookie= sr.ReadToEnd();
sr.Close();
}
//string tmp = SRWebClient.GetPage("http://bbs.chinaz.com/Members.html?page=1&sort=CreateDate&desc=true&keyword=", Encoding.UTF8, cookie);
int a =int.Parse(Console.ReadLine());
int b =int.Parse(Console.ReadLine());
string url = Console.ReadLine();
Hashtable hash=new Hashtable();
Encoding encoding= Encoding.GetEncoding(Console.ReadLine());
for (int i = a; i <= b; i++)
{
string html = SRWebClient.GetPage(string.Format(url, i), encoding, cookie);
//Console.WriteLine(html);
if (html !=null&& html.Length >1000)
{
Match m= Regex.Match(html, @"/w+([-+.']/w+)*@/w+([-.]/w+)*/https://www.cnblogs.com/szgzwf/archive/2012/11/05/w+([-.]/w+)*", RegexOptions.Compiled | RegexOptions.IgnoreCase);
while (m !=null&& m.Value !=null&& m.Value.Trim() !=string.Empty)
{
if (!hash.Contains(m.Value))
{
Console.WriteLine(m.Value);
using (StreamWriter sw =new StreamWriter("mail.txt", true))
{
sw.WriteLine(m.Value);
sw.Close();
}
hash.Add(m.Value,string.Empty);
}
m= m.NextMatch();
}
}
}
Console.Write("完成");
Console.ReadLine();
}
}
publicclass SRWebClient
{
public cookieCollection cookie;
public SRWebClient()
{
cookie =null;
}
#region 从包含多个 cookie 的字符串读取到 cookieCollection 集合中
privatestaticvoid AddcookieWithcookieHead(ref cookieCollection cookieCol, string cookieHead, string defaultDomain)
{
if (cookieCol ==null) cookieCol =new cookieCollection();
if (cookieHead ==null) return;
string[] ary = cookieHead.Split(';');
for (int i =0; i < ary.Length; i++)
{
cookie ck = GetcookieFromString(ary[i].Trim(), defaultDomain);
if (ck !=null)
{
cookieCol.Add(ck);
}
}
}
#endregion
#region 读取某一个 cookie 字符串到 cookie 变量中
privatestatic cookie GetcookieFromString(string cookieString, string
{
string[] ary = cookieString.Split(',');
Hashtable hs =new Hashtable();
for (int i =0; i < ary.Length; i++)
{
string s = ary[i].Trim();
int index = s.IndexOf("=");
if (index >0)
{
hs.Add(s.Substring(0, index), s.Substring(index +1));
}
}
cookie ck =new cookie();
foreach (object Key in hs.Keys)
{
if (Key.ToString() =="path") ck.Path = hs[Key].ToString();
elseif (Key.ToString() =="expires")
{
//ck.Expires=DateTime.Parse(hs[Key].ToString();
}
elseif (Key.ToString() =="domain") ck.Domain = hs[Key].ToString();
else
{
ck.Name = Key.ToString();
ck.Value = hs[Key].ToString();
}
}
if (ck.Name =="") returnnull;
if (ck.Domain =="") ck.Domain = defaultDomain;
return ck;
}
#endregion
///<TgData>
///<Alias>下载Web源代码</Alias>
///</TgData>
publicstring DownloadHtml(string URL, bool Createcookie)
{
try
{
HttpWebRequest request = HttpWebRequest.Create(URL) as HttpWebRequest;
if (cookie !=null)
{
request.cookieContainer =new cookieContainer();
request.cookieContainer.Add(cookie);
}
request.AllowAutoRedirect =false;
//request.MaximumAutomaticRedirections = 3;
request.Timeout =20000;
HttpWebResponse res = (HttpWebResponse)request.GetResponse();
string r ="";
System.IO.StreamReader S1 =new System.IO.StreamReader(res.GetResponseStream(), System.Text.Encoding.Default);
try
{
r = S1.ReadToEnd();
if (Createcookie)
cookie = res.cookies;
}
catch (Exception er)
{
//Log l = new Log();
//l.writelog("下载Web错误", er.ToString());
}
finally
{
res.Close();
S1.Close();
}
return r;
}
catch
{
}
returnstring.Empty;
}
以上就是本篇文章【正则表达式提取链接地址-邮箱邮件】的全部内容了，欢迎阅览 ！ 文章地址：http://www78564.xrbh.cn/quote/27361.html
动态
相关文章
文章
同类文章
热门文章
栏目首页
网站地图
返回首页 迅博思语移动站 http://www78564.xrbh.cn/mobile/ , 查看更多