private HtmlNodeNavigator(HtmlNodeNavigator nav)
{
if (nav == null)
{
throw new ArgumentNullException("nav");
}
InternalTrace(null);
_doc = nav._doc;
_currentnode = nav._currentnode;
_attindex = nav._attindex;
_nametable = nav._nametable; // REVIEW: should we do this?
}
internal HtmlNodeNavigator(HtmlDocument doc, HtmlNode currentNode)
{
if (currentNode == null)
{
throw new ArgumentNullException("currentNode");
}
if (currentNode.OwnerDocument != doc)
{
throw new ArgumentException(HtmlDocument.HtmlExceptionRefNotChild);
}
InternalTrace(null);
_doc = doc;
Reset();
_currentnode = currentNode;
}
/// <summary>
/// Creates an HTML node from a string representing literal HTML.
/// </summary>
/// <param name="html">The HTML text.</param>
/// <returns>The newly created node instance.</returns>
public static HtmlNode CreateNode(string html)
{
// REVIEW: this is *not* optimum...
HtmlDocument doc = new HtmlDocument();
doc.LoadHtml(html);
return doc.DocumentNode.FirstChild;
}
/// <summary>
/// Initializes HtmlNode, providing type, owner and where it exists in a collection
/// </summary>
/// <param name="type"></param>
/// <param name="ownerdocument"></param>
/// <param name="index"></param>
public HtmlNode(HtmlNodeType type, HtmlDocument ownerdocument, int index)
{
_nodetype = type;
_ownerdocument = ownerdocument;
_outerstartindex = index;
switch (type)
{
case HtmlNodeType.Comment:
Name = HtmlNodeTypeNameComment;
_endnode = this;
break;
case HtmlNodeType.Document:
Name = HtmlNodeTypeNameDocument;
_endnode = this;
break;
case HtmlNodeType.Text:
Name = HtmlNodeTypeNameText;
_endnode = this;
break;
}
if (_ownerdocument.Openednodes != null)
{
if (!Closed)
{
// we use the index as the key
// -1 means the node comes from public
if (-1 != index)
{
_ownerdocument.Openednodes.Add(index, this);
}
}
}
if ((-1 != index) || (type == HtmlNodeType.Comment) || (type == HtmlNodeType.Text)) return;
// innerhtml and outerhtml must be calculated
SetChanged();
}
private HtmlDocument LoadUrl(Uri uri, string method, WebProxy proxy, NetworkCredential creds)
{
HtmlDocument doc = new HtmlDocument();
doc.OptionAutoCloseOnEnd = false;
doc.OptionFixNestedTags = true;
_statusCode = Get(uri, method, null, doc, proxy, creds);
if (_statusCode == HttpStatusCode.NotModified)
{
// read cached encoding
doc.DetectEncodingAndLoad(GetCachePath(uri));
}
return doc;
}
private HttpStatusCode Get(Uri uri, string method, string path, HtmlDocument doc, IWebProxy proxy,
ICredentials creds)
{
string cachePath = null;
HttpWebRequest req;
bool oldFile = false;
req = WebRequest.Create(uri) as HttpWebRequest;
req.Method = method;
req.UserAgent = UserAgent;
if (proxy != null)
{
if (creds != null)
{
proxy.Credentials = creds;
req.Credentials = creds;
}
else
{
proxy.Credentials = CredentialCache.DefaultCredentials;
req.Credentials = CredentialCache.DefaultCredentials;
}
req.Proxy = proxy;
}
_fromCache = false;
_requestDuration = 0;
int tc = Environment.TickCount;
if (UsingCache)
{
cachePath = GetCachePath(req.RequestUri);
if (File.Exists(cachePath))
{
req.IfModifiedSince = File.GetLastAccessTime(cachePath);
oldFile = true;
}
}
if (_cacheOnly)
{
if (!File.Exists(cachePath))
{
throw new HtmlWebException("File was not found at cache path: '" + cachePath + "'");
}
if (path != null)
{
IOLibrary.CopyAlways(cachePath, path);
// touch the file
if (cachePath != null) File.SetLastWriteTime(path, File.GetLastWriteTime(cachePath));
}
_fromCache = true;
return HttpStatusCode.NotModified;
}
if (_useCookies)
{
req.CookieContainer = new CookieContainer();
}
if (PreRequest != null)
{
// allow our user to change the request at will
if (!PreRequest(req))
{
return HttpStatusCode.ResetContent;
}
// dump cookie
// if (_useCookies)
// {
// foreach(Cookie cookie in req.CookieContainer.GetCookies(req.RequestUri))
// {
// HtmlLibrary.Trace("Cookie " + cookie.Name + "=" + cookie.Value + " path=" + cookie.Path + " domain=" + cookie.Domain);
// }
// }
}
HttpWebResponse resp;
try
{
resp = req.GetResponse() as HttpWebResponse;
}
catch (WebException we)
{
_requestDuration = Environment.TickCount - tc;
resp = (HttpWebResponse)we.Response;
if (resp == null)
{
if (oldFile)
{
if (path != null)
{
IOLibrary.CopyAlways(cachePath, path);
// touch the file
File.SetLastWriteTime(path, File.GetLastWriteTime(cachePath));
}
return HttpStatusCode.NotModified;
}
throw;
}
}
catch (Exception)
{
_requestDuration = Environment.TickCount - tc;
throw;
}
// allow our user to get some info from the response
if (PostResponse != null)
{
PostResponse(req, resp);
}
_requestDuration = Environment.TickCount - tc;
_responseUri = resp.ResponseUri;
bool html = IsHtmlContent(resp.ContentType);
Encoding respenc = !string.IsNullOrEmpty(resp.ContentEncoding)
? Encoding.GetEncoding(resp.ContentEncoding)
: null;
if (OverrideEncoding != null)
respenc = OverrideEncoding;
if (resp.StatusCode == HttpStatusCode.NotModified)
{
if (UsingCache)
{
_fromCache = true;
if (path != null)
{
IOLibrary.CopyAlways(cachePath, path);
// touch the file
File.SetLastWriteTime(path, File.GetLastWriteTime(cachePath));
}
return resp.StatusCode;
}
// this should *never* happen...
throw new HtmlWebException("Server has send a NotModifed code, without cache enabled.");
}
Stream s = resp.GetResponseStream();
if (s != null)
{
if (UsingCache)
{
// NOTE: LastModified does not contain milliseconds, so we remove them to the file
SaveStream(s, cachePath, RemoveMilliseconds(resp.LastModified), _streamBufferSize);
// save headers
SaveCacheHeaders(req.RequestUri, resp);
if (path != null)
{
// copy and touch the file
IOLibrary.CopyAlways(cachePath, path);
File.SetLastWriteTime(path, File.GetLastWriteTime(cachePath));
}
}
else
{
// try to work in-memory
if (doc != null && html)
{
if (respenc == null)
{
doc.Load(s, true);
}
else
{
doc.Load(s, respenc);
}
}
}
resp.Close();
}
return resp.StatusCode;
}
/// <summary>
/// Loads an HTML document from an Internet resource.
/// </summary>
/// <param name="url">The requested URL, such as "http://Myserver/Mypath/Myfile.asp".</param>
/// <param name="method">The HTTP method used to open the connection, such as GET, POST, PUT, or PROPFIND.</param>
/// <param name="proxy">Proxy to use with this request</param>
/// <param name="credentials">Credentials to use when authenticating</param>
/// <returns>A new HTML document.</returns>
public HtmlDocument Load(string url, string method, WebProxy proxy, NetworkCredential credentials)
{
Uri uri = new Uri(url);
HtmlDocument doc;
if ((uri.Scheme == Uri.UriSchemeHttps) ||
(uri.Scheme == Uri.UriSchemeHttp))
{
doc = LoadUrl(uri, method, proxy, credentials);
}
else
{
if (uri.Scheme == Uri.UriSchemeFile)
{
doc = new HtmlDocument();
doc.OptionAutoCloseOnEnd = false;
doc.OptionAutoCloseOnEnd = true;
doc.DetectEncodingAndLoad(url, _autoDetectEncoding);
}
else
{
throw new HtmlWebException("Unsupported uri scheme: '" + uri.Scheme + "'.");
}
}
if (PreHandleDocument != null)
{
PreHandleDocument(doc);
}
return doc;
}
internal HtmlAttribute(HtmlDocument ownerdocument)
{
_ownerdocument = ownerdocument;
}
/// <summary>
/// 得到安全的Html
/// </summary>
/// <param name="html"></param>
/// <returns></returns>
public static string GetSafeHtml(string html)
{
if (html.IsNullOrEmpty())
return html;
HtmlDocument doc = new HtmlDocument();
doc.LoadHtml(html);
List<HtmlNode> nodesToRemove = new List<HtmlNode>();
IEnumerator e = doc
.CreateNavigator()
.SelectDescendants(System.Xml.XPath.XPathNodeType.All, false)
.GetEnumerator();
while (e.MoveNext())
{
HtmlNode node =
((HtmlNodeNavigator)e.Current)
.CurrentNode;
if (_DefaultNodeBlackList.Contains(node.Name))
nodesToRemove.Add(node);
else
{
FilterAttributes(node);
FilterScriptRelAttributes(node);
}
}
nodesToRemove.ForEach(node => node.ParentNode.RemoveChild(node));
StringBuilder strB = new StringBuilder();
using (StringWriter writer = new StringWriter(strB))
doc.Save(writer);
return strB.ToString();
}
internal HtmlCommentNode(HtmlDocument ownerdocument, int index)
:
base(HtmlNodeType.Comment, ownerdocument, index)
{
}
