/// <summary>
/// 根据相传入的数据，得到相应页面数据
/// </summary>
/// <param name="item">参数类对象</param>
/// <returns>返回HttpResult类型</returns>
public HttpResult GetHtml(HttpItem item)
{
//返回参数
HttpResult result = new HttpResult();
try
{
//准备参数
SetRequest(item);
}
catch (Exception ex)
{
result.Cookie = string.Empty;
result.Header = null;
result.Html = ex.Message;
result.StatusDescription = "配置参数时出错：" + ex.Message;
//配置参数时出错
return result;
}
try
{
//请求数据
using (response = (HttpWebResponse)request.GetResponse())
{
GetData(item, result);
}
}
catch (WebException ex)
{
if (ex.Response != null)
{
using (response = (HttpWebResponse)ex.Response)
{
GetData(item, result);
}
}
else
{
result.Html = ex.Message;
}
}
catch (Exception ex)
{
result.Html = ex.Message;
}
if (item.IsToLower) result.Html = result.Html.ToLower();
return result;
}
/// <summary>
/// 获取数据的并解析的方法
/// </summary>
/// <param name="item"></param>
/// <param name="result"></param>
private void GetData(HttpItem item, HttpResult result)
{
#region base
//获取StatusCode
result.StatusCode = response.StatusCode;
//获取StatusDescription
result.StatusDescription = response.StatusDescription;
//获取Headers
result.Header = response.Headers;
//获取CookieCollection
if (response.Cookies != null) result.CookieCollection = response.Cookies;
//获取set-cookie
if (response.Headers["set-cookie"] != null) result.Cookie = response.Headers["set-cookie"];
#endregion
#region byte
//处理网页Byte
byte[] ResponseByte = GetByte();
#endregion
#region Html
if (ResponseByte != null & ResponseByte.Length > 0)
{
//设置编码
SetEncoding(item, result, ResponseByte);
//得到返回的HTML
result.Html = encoding.GetString(ResponseByte);
}
else
{
//没有返回任何Html代码
result.Html = string.Empty;
}
#endregion
}
/// <summary>
/// 设置编码
/// </summary>
/// <param name="item">HttpItem</param>
/// <param name="result">HttpResult</param>
/// <param name="ResponseByte">byte[]</param>
private void SetEncoding(HttpItem item, HttpResult result, byte[] ResponseByte)
{
//是否返回Byte类型数据
if (item.ResultType == ResultType.Byte) result.ResultByte = ResponseByte;
//从这里开始我们要无视编码了
if (encoding == null)
{
Match meta = Regex.Match(Encoding.Default.GetString(ResponseByte), "<meta[^<]*charset=([^<]*)[\"']", RegexOptions.IgnoreCase);
string c = string.Empty;
if (meta != null && meta.Groups.Count > 0)
{
c = meta.Groups[1].Value.ToLower().Trim();
}
if (c.Length > 2)
{
try
{
encoding = Encoding.GetEncoding(c.Replace("\"", string.Empty).Replace("'", "").Replace(";", "").Replace("iso-8859-1", "gbk").Trim());
}
catch
{
if (string.IsNullOrEmpty(response.CharacterSet))
{
encoding = Encoding.UTF8;
}
else
{
encoding = Encoding.GetEncoding(response.CharacterSet);
}
}
}
else
{
if (string.IsNullOrEmpty(response.CharacterSet))
{
encoding = Encoding.UTF8;
}
else
{
encoding = Encoding.GetEncoding(response.CharacterSet);
}
}
}
}