private void TraverseNode(HtmlNode htmlNode)
{
if (htmlNode == null || htmlNode.NodeType == HtmlNodeType.Comment)
{
return;
}
var attributes = new Attributes();
if (htmlNode.HasAttributes)
{
foreach (HtmlAttribute attribute in htmlNode.Attributes)
{
attributes.AddAttribute(null, htmlNode.Name, attribute.Name, null, attribute.Value);
}
}
ContentHandler.StartElement(null, htmlNode.Name, htmlNode.Name, attributes);
if (htmlNode.NodeType == HtmlNodeType.Text)
{
ContentHandler.Characters(htmlNode.InnerText.ToCharArray(), 0, htmlNode.InnerText.Length);
}
else if (htmlNode.HasChildNodes)
{
foreach (HtmlNode childNode in htmlNode.ChildNodes)
{
TraverseNode(childNode);
}
}
ContentHandler.EndElement(null, htmlNode.Name, htmlNode.Name);
}
/// <summary>Add a SAX tag with a string inside.</summary>
/// <param name="contentHandler">the SAX content handler</param>
/// <param name="tag">the element tag to use</param>
/// <param name="val">the string to put inside the tag</param>
/// <exception cref="Org.Xml.Sax.SAXException"/>
public static void AddSaxString(ContentHandler contentHandler, string tag, string
val)
{
contentHandler.StartElement(string.Empty, string.Empty, tag, new AttributesImpl()
);
char[] c = MangleXmlString(val, false).ToCharArray();
contentHandler.Characters(c, 0, c.Length);
contentHandler.EndElement(string.Empty, string.Empty, tag);
}
/// <summary>Start visitor (initialization)</summary>
/// <exception cref="System.IO.IOException"/>
public virtual void Start(int version)
{
try
{
contentHandler.StartElement(string.Empty, string.Empty, "EDITS_VERSION", new AttributesImpl
());
StringBuilder bld = new StringBuilder();
bld.Append(version);
AddString(bld.ToString());
contentHandler.EndElement(string.Empty, string.Empty, "EDITS_VERSION");
}
catch (SAXException e)
{
throw new IOException("SAX error: " + e.Message);
}
}
/// <summary>
/// Create a processor that writes to the file named and may or may not
/// also output to the screen, as specified.
/// </summary>
/// <param name="filename">Name of file to write output to</param>
/// <param name="printToScreen">Mirror output to screen?</param>
/// <exception cref="System.IO.IOException"/>
public XmlEditsVisitor(OutputStream @out)
{
this.@out = @out;
OutputFormat outFormat = new OutputFormat("XML", "UTF-8", true);
outFormat.SetIndenting(true);
outFormat.SetIndent(2);
outFormat.SetDoctype(null, null);
XMLSerializer serializer = new XMLSerializer(@out, outFormat);
contentHandler = serializer.AsContentHandler();
try
{
contentHandler.StartDocument();
contentHandler.StartElement(string.Empty, string.Empty, "EDITS", new AttributesImpl
());
}
catch (SAXException e)
{
throw new IOException("SAX error: " + e.Message);
}
}
public void Parse(StreamReader input, ContentHandler handler)
{
var buffer = new StringBuilder();
var isInsideTag = false;
var isStartTag = true;
var lastChar = -1;
int c;
while ((c = input.Read()) != -1)
{
if ('<' == c)
{
if (isInsideTag)
{
throw new InvalidFormatException("Did not expect < char!");
}
if (buffer.ToString().Trim().Length > 0)
{
handler.Characters(buffer.ToString().Trim());
}
buffer = new StringBuilder();
isInsideTag = true;
isStartTag = true;
}
// buffer.AppendCodePoint(c); <-- java
buffer.Append(char.ConvertFromUtf32(c));
if ('/' == c && lastChar == '<')
{
isStartTag = false;
}
if ('>' == c)
{
if (!isInsideTag)
{
throw new InvalidFormatException("Did not expect > char!");
}
if (isStartTag)
{
handler.StartElement(extractTagName(buffer), getAttributes(buffer));
}
else
{
handler.EndElement(extractTagName(buffer));
}
buffer = new StringBuilder();
isInsideTag = false;
}
lastChar = c;
}
if (isInsideTag)
{
throw new InvalidFormatException("Did not find matching > char!");
}
}
public void Parse(StreamReader input, ContentHandler handler) {
var buffer = new StringBuilder();
var isInsideTag = false;
var isStartTag = true;
var lastChar = -1;
int c;
while ((c = input.Read()) != -1) {
if ('<' == c) {
if (isInsideTag) {
throw new InvalidFormatException("Did not expect < char!");
}
if (buffer.ToString().Trim().Length > 0) {
handler.Characters(buffer.ToString().Trim());
}
buffer = new StringBuilder();
isInsideTag = true;
isStartTag = true;
}
// buffer.AppendCodePoint(c); <-- java
buffer.Append(char.ConvertFromUtf32(c));
if ('/' == c && lastChar == '<') {
isStartTag = false;
}
if ('>' == c) {
if (!isInsideTag) {
throw new InvalidFormatException("Did not expect > char!");
}
if (isStartTag) {
handler.StartElement(extractTagName(buffer), getAttributes(buffer));
} else {
handler.EndElement(extractTagName(buffer));
}
buffer = new StringBuilder();
isInsideTag = false;
}
lastChar = c;
}
if (isInsideTag) {
throw new InvalidFormatException("Did not find matching > char!");
}
}
private void TraverseNode(HtmlNode htmlNode)
{
if (htmlNode.NodeType == HtmlNodeType.Element)
{
var attributes = new Attributes();
if (htmlNode.HasAttributes)
{
foreach (HtmlAttribute attribute in htmlNode.Attributes)
{
attributes.AddAttribute(null, htmlNode.Name, attribute.Name, null, attribute.Value);
}
}
//Debug.WriteLine($"START: {htmlNode.Name}");
ContentHandler.StartElement(null, htmlNode.Name, htmlNode.Name, attributes);
if (htmlNode.HasChildNodes)
{
foreach (HtmlNode childNode in htmlNode.ChildNodes)
{
if (childNode.NodeType == HtmlNodeType.Text)
{
//Debug.WriteLine(childNode.InnerText);
ContentHandler.Characters(childNode.InnerText.ToCharArray(), 0, childNode.InnerText.Length);
}
TraverseNode(childNode);
}
}
//Debug.WriteLine($"ENDE: {htmlNode.Name}");
ContentHandler.EndElement(null, htmlNode.Name, htmlNode.Name);
}
else
{
if (htmlNode.HasChildNodes)
{
foreach (HtmlNode childNode in htmlNode.ChildNodes)
{
TraverseNode(childNode);
}
}
}
//if (!(htmlNode.NodeType == HtmlNodeType.Comment ||
// htmlNode.NodeType == HtmlNodeType.Text ||
// htmlNode.NodeType == HtmlNodeType.))
//{
// ContentHandler.StartElement(null, htmlNode.Name, htmlNode.Name, null);
// if (htmlNode.HasChildNodes)
// foreach (HtmlNode childNode in htmlNode.ChildNodes)
// TraverseNode(childNode);
// ContentHandler.EndElement(null, htmlNode.Name, htmlNode.Name);
//}
//else
//{
// if (htmlNode.HasChildNodes)
// foreach (HtmlNode childNode in htmlNode.ChildNodes)
// TraverseNode(childNode);
//}
}