asp.net HttpWebRequest自动识别网页编码

所属分类: 网络编程 / ASP.NET 阅读数: 1316
收藏 0 赞 0 分享

复制代码 代码如下:

static string GetEncoding(string url)
{
HttpWebRequest request = null;
HttpWebResponse response = null;
StreamReader reader = null;
try
{
request = (HttpWebRequest)WebRequest.Create(url);
request.Timeout = 20000;
request.AllowAutoRedirect = false;

response = (HttpWebResponse)request.GetResponse();
if (response.StatusCode == HttpStatusCode.OK && response.ContentLength < 1024 * 1024)
{
if (response.ContentEncoding != null && response.ContentEncoding.Equals("gzip", StringComparison.InvariantCultureIgnoreCase))
reader = new StreamReader(new GZipStream(response.GetResponseStream(), CompressionMode.Decompress));
else
reader = new StreamReader(response.GetResponseStream(), Encoding.ASCII);

string html = reader.ReadToEnd();

Regex reg_charset = new Regex(@"charset\b\s*=\s*(?<charset>[^""]*)");
if (reg_charset.IsMatch(html))
{
return reg_charset.Match(html).Groups["charset"].Value;
}
else if (response.CharacterSet != string.Empty)
{
return response.CharacterSet;
}
else
return Encoding.Default.BodyName;
}
}
catch
{
}
finally
{

if (response != null)
{
response.Close();
response = null;
}
if (reader != null)
reader.Close();

if (request != null)
request = null;

}

return Encoding.Default.BodyName;
}

/// <summary>
/// 获取源代码
/// </summary>
/// <param name="url"></param>
/// <returns></returns>
static string GetHtml(string url, Encoding encoding)
{
HttpWebRequest request = null;
HttpWebResponse response = null;
StreamReader reader = null;
try
{
request = (HttpWebRequest)WebRequest.Create(url);
request.Timeout = 20000;
request.AllowAutoRedirect = false;

response = (HttpWebResponse)request.GetResponse();
if (response.StatusCode == HttpStatusCode.OK && response.ContentLength < 1024 * 1024)
{
if (response.ContentEncoding != null && response.ContentEncoding.Equals("gzip", StringComparison.InvariantCultureIgnoreCase))
reader = new StreamReader(new GZipStream(response.GetResponseStream(), CompressionMode.Decompress), encoding);
else
reader = new StreamReader(response.GetResponseStream(), encoding);
string html = reader.ReadToEnd();

return html;
}
}
catch
{
}
finally
{

if (response != null)
{
response.Close();
response = null;
}
if (reader != null)
reader.Close();

if (request != null)
request = null;

}

return string.Empty;
}

更多精彩内容其他人还在看

Asp.net中的mail的发送

Asp.net中的mail的发送
收藏 0 赞 0 分享

用ASP.Net实现文件的在线压缩和解压缩

用ASP.Net实现文件的在线压缩和解压缩
收藏 0 赞 0 分享

ASP.NET中文件上传下载方法集合

ASP.NET中文件上传下载方法集合
收藏 0 赞 0 分享

ASP.NET通过Remoting service上传文件

ASP.NET通过Remoting service上传文件
收藏 0 赞 0 分享

ASP.NET2.0服务器控件之Render方法

ASP.NET2.0服务器控件之Render方法
收藏 0 赞 0 分享

ASP.NET2.0 WebRource,开发微调按钮控件

ASP.NET2.0 WebRource,开发微调按钮控件
收藏 0 赞 0 分享

ASP.NET2.0新特性概述

ASP.NET2.0新特性概述
收藏 0 赞 0 分享

介绍几个ASP.NET中容易忽略但却很重要的方法函数

介绍几个ASP.NET中容易忽略但却很重要的方法函数
收藏 0 赞 0 分享

asp.net2.0如何加密数据库联接字符串

asp.net2.0如何加密数据库联接字符串
收藏 0 赞 0 分享

用.NET 2.0压缩/解压功能处理大型数据

用.NET 2.0压缩/解压功能处理大型数据
收藏 0 赞 0 分享
查看更多