黄聪:C#中WebClient自动判断编码是UTF-8还是GBK,并且有超时判断功能

时间:2023-03-08 22:16:04
public class WebDownload : WebClient
{
private int _timeout;
/// <summary>
/// 超时时间(毫秒)
/// </summary>
public int Timeout
{
get
{
return _timeout;
}
set
{
_timeout = value;
}
} public WebDownload()
{
this._timeout = ;
} public WebDownload(int timeout)
{
this._timeout = timeout;
} protected override WebRequest GetWebRequest(Uri address)
{
var result = base.GetWebRequest(address);
result.Timeout = this._timeout;
return result;
} public string Get(string url)
{
string html = "";
var data = this.DownloadData(url);
var r_utf8 = new System.IO.StreamReader(new System.IO.MemoryStream(data), Encoding.UTF8); //将html放到utf8编码的StreamReader内
var r_gbk = new System.IO.StreamReader(new System.IO.MemoryStream(data), Encoding.Default); //将html放到gbk编码的StreamReader内
var t_utf8 = r_utf8.ReadToEnd(); //读出html内容
var t_gbk = r_gbk.ReadToEnd(); //读出html内容
if (!isLuan(t_utf8)) //判断utf8是否有乱码
{
html = t_utf8;
}
else
{
html = t_gbk;
} return html;
} bool isLuan(string txt)
{
var bytes = Encoding.UTF8.GetBytes(txt);
//239 191 189
for (var i = ; i < bytes.Length; i++)
{
if (i < bytes.Length - )
if (bytes[i] == && bytes[i + ] == && bytes[i + ] == )
{
return true;
}
}
return false;
} }