public void CompareLowerCulture()
{
string html = File.ReadAllText(Path.Combine(_contentDirectory, "regression.html"));
HtmlNode node1 = null;
// Test 1
CultureInfo cul1 = CultureInfo.CreateSpecificCulture("en-US");
Thread.CurrentThread.CurrentCulture = cul1;
Candy.HtmlDocument doc1 = new Candy.HtmlDocument();
doc1.LoadHtml(html);
node1 = doc1.DocumentNode.SelectSingleNode("//div[@id='mainContents']/h2");
CultureInfo cul2 = CultureInfo.CreateSpecificCulture("tr-TR");
Thread.CurrentThread.CurrentCulture = cul2;
Candy.HtmlDocument doc2 = new Candy.HtmlDocument();
doc2.LoadHtml(html);
var s = doc2.DocumentNode.OuterHtml;
HtmlNode node2 = doc2.DocumentNode.SelectSingleNode("//div[@id='mainContents']/h2");
if (node1?.InnerHtml == node2?.InnerHtml)
{
Assert.AreEqual(node1?.InnerHtml, node2?.InnerHtml);
}
Assert.AreEqual(0, doc2.DocumentNode.OwnerDocument.ParseErrors.Count());
}
public void TestCommentNode()
{
var html =
@"<!DOCTYPE html>
<html>
<body>
<!--title='Title' >
<!--title='Title'--!>
<!--title = 'Title'-->
<!--title='Title'--!>
<h1>This is <b>bold</b> headddding</h1>
<p>This is <u>underlinyed</u> paragraph</p>
<h2>This is <i>italic</i> heading</h2>
</body>
</html> ";
var htmlDoc = new Candy.HtmlDocument();
htmlDoc.LoadHtml(html);
var h1 = htmlDoc.DocumentNode.SelectNodes("//h1");
var comments = htmlDoc.DocumentNode.SelectNodes("//comment()");
Assert.AreEqual(h1.Count, 1);
Assert.AreEqual(comments.Count, 4);
}
public void ChangesToPHandlingFalse()
{
HtmlDocument.DisableBehaviorTagP = false;
var input = "<p>Begin<div>Inner</div>End</p>";
var doc = new Candy.HtmlDocument();
doc.LoadHtml(input);
var docNode = doc.DocumentNode;
try
{
// Check tree is correct
Assert.AreEqual(1, docNode.ChildNodes.Count);
Assert.AreEqual("p", docNode.ChildNodes[0].Name);
var pRootNode = docNode.ChildNodes[0];
Assert.AreEqual(3, pRootNode.ChildNodes.Count);
Assert.AreEqual("Begin", pRootNode.ChildNodes[0].InnerText);
Assert.AreEqual("Inner", pRootNode.ChildNodes[1].InnerText);
Assert.AreEqual("End", pRootNode.ChildNodes[2].InnerText);
}
catch (Exception e)
{
throw;
}
finally
{
HtmlDocument.DisableBehaviorTagP = true;
}
}
public void checkAttributForTextComment()
{
var doc = new Candy.HtmlDocument();
doc.LoadHtml(@"<html><body><div id='foo'><span> some</span> text</div></body></html>");
var div = doc.GetElementbyId("foo");
int count = 0;
Exception exception = null;
foreach (var textNode in div.ChildNodes)
{
try
{
textNode.Id = "1";
count++;
}
catch (Exception e)
{
exception = e;
}
}
Assert.AreEqual(count, 1);
Assert.IsNotNull(exception);
}
public void TestEmptyTag_Single()
{
var html = "<img src=\"x\"/onerror=\"alert('onerror1')\"><img/src=\"x\"/onerror=\"alert('onerror2')\">";
var doc = new Candy.HtmlDocument();
doc.LoadHtml(html);
Assert.AreEqual(@"<img src=""x"" onerror=""alert('onerror1')""><img src=""x"" onerror=""alert('onerror2')"">", doc.DocumentNode.OuterHtml);
}
public void ScriptingText()
{
var html = @"<?xml version=""1.0"" encoding=""UTF-8"" ?>
<html xmlns=""http://www.w3.org/1999/xhtml"">
<head>
<title>SEE title</title>
<script>SEE script </script>
<style>SEE style</style>
</head>
<body>
<script>NOTSEE script</script>
<div>222<script>NOTSEE script</script>
<style>NOTSEE style</style></div>
</body>
</html>";
{
Candy.HtmlDocument htmlDocument = new Candy.HtmlDocument();
htmlDocument.LoadHtml(html);
var content1 = htmlDocument.DocumentNode.SelectSingleNode("//head").InnerText;
var content2 = htmlDocument.DocumentNode.SelectSingleNode("//script").InnerText;
var content3 = htmlDocument.DocumentNode.SelectSingleNode("//style").InnerText;
var content4 = htmlDocument.DocumentNode.SelectSingleNode("//body").InnerText;
var content5 = htmlDocument.DocumentNode.SelectSingleNode("//html").InnerText;
var content6 = htmlDocument.DocumentNode.SelectSingleNode("//body/script").InnerText;
Assert.AreEqual("\r\n SEE title\r\n\tSEE script \r\n\tSEE style\r\n", content1);
Assert.AreEqual("SEE script ", content2);
Assert.AreEqual("SEE style", content3);
Assert.AreEqual("\r\n\r\n222\r\n\r\n", content4);
Assert.AreEqual("\r\n\r\n SEE title\r\n\t\r\n\t\r\n\r\n\r\n\r\n222\r\n\r\n\r\n", content5);
Assert.AreEqual("NOTSEE script", content6);
}
{
Candy.HtmlDocument htmlDocument = new Candy.HtmlDocument();
htmlDocument.BackwardCompatibility = false;
htmlDocument.LoadHtml(html);
var content1 = htmlDocument.DocumentNode.SelectSingleNode("//head").InnerText;
var content2 = htmlDocument.DocumentNode.SelectSingleNode("//script").InnerText;
var content3 = htmlDocument.DocumentNode.SelectSingleNode("//style").InnerText;
var content4 = htmlDocument.DocumentNode.SelectSingleNode("//body").InnerText;
var content5 = htmlDocument.DocumentNode.SelectSingleNode("//html").InnerText;
var content6 = htmlDocument.DocumentNode.SelectSingleNode("//body/script").InnerText;
Assert.AreEqual(" SEE titleSEE script SEE style", content1);
Assert.AreEqual("SEE script ", content2);
Assert.AreEqual("SEE style", content3);
Assert.AreEqual("222", content4);
Assert.AreEqual(" SEE title222", content5);
Assert.AreEqual("NOTSEE script", content6);
}
}
public void testTEXTAREA()
{
{
var doc = new Candy.HtmlDocument();
doc.LoadHtml(@"<script><div>hello</div></script><TEXTAREA>Text in the <div>hello</div>area</TEXTAREA>");
HtmlNodeCollection divs = doc.DocumentNode.SelectNodes("//div");
Assert.IsNull(divs);
HtmlNode ta = doc.DocumentNode.SelectSingleNode("//textarea");
Assert.IsTrue(ta.InnerHtml.Contains("div"));
}
}
public void OverFlowNotEndTag()
{
string html = File.ReadAllText(Path.Combine(_contentDirectory, "overflow.html"));
HtmlNode node1 = null;
// Test 1
Candy.HtmlDocument doc1 = new Candy.HtmlDocument();
doc1.LoadHtml(html);
Assert.AreEqual(15, doc1.DocumentNode.ChildNodes[4].ChildNodes.Count);
Assert.AreEqual(0, doc1.DocumentNode.OwnerDocument.ParseErrors.Count());
}
public void ChangesToPHandling()
{
var input = "<p>Begin<div>Inner</div>End</p>";
var doc = new Candy.HtmlDocument();
doc.LoadHtml(input);
var docNode = doc.DocumentNode;
// Check tree is correct
Assert.AreEqual(3, docNode.ChildNodes.Count);
Assert.AreEqual("p", docNode.ChildNodes[0].Name);
Assert.AreEqual("div", docNode.ChildNodes[1].Name);
Assert.AreEqual("#text", docNode.ChildNodes[2].Name);
Assert.AreEqual("Begin", docNode.ChildNodes[0].InnerText);
Assert.AreEqual("Inner", docNode.ChildNodes[1].InnerText);
Assert.AreEqual("End", docNode.ChildNodes[2].InnerText);
}
public void SanitizeXmlElementNameWithColon()
{
var input = @"<RootElement xmlns:MyNamespace=""value"">
<value:element />
</RootElement>";
var htmlDoc = new Candy.HtmlDocument();
htmlDoc.LoadHtml(input);
htmlDoc.OptionDefaultStreamEncoding = System.Text.Encoding.UTF8;
htmlDoc.OptionOutputAsXml = true;
htmlDoc.OptionOutputOriginalCase = true;
var xmlDoc = htmlDoc.DocumentNode.WriteTo();
var expected = @"<?xml version=""1.0"" encoding=""utf-8""?>" +
@"<RootElement xmlns:MyNamespace=""value"">
<_value3a_element></_value3a_element>
</RootElement>";
Assert.AreEqual(expected, xmlDoc);
}
public void TextInsideScriptTagShouldHaveCorrectStreamPosition()
{
{
var document = new HtmlDocument();
document.LoadHtml(@"<scrapt>foo</scrapt>");
var scraptText = document.DocumentNode.FirstChild.FirstChild;
Assert.AreEqual(8, scraptText.StreamPosition);
Assert.AreEqual(1, scraptText.Line);
Assert.AreEqual(8, scraptText.LinePosition);
}
{
var document = new HtmlDocument();
document.LoadHtml(@"<script>foo</script>");
var scriptText = document.DocumentNode.FirstChild.FirstChild;
Assert.AreEqual(8, scriptText.StreamPosition);
Assert.AreEqual(1, scriptText.Line);
Assert.AreEqual(8, scriptText.LinePosition);
}
{
var document = new Candy.HtmlDocument();
document.LoadHtml(@"
<scrapt>foo</scrapt>");
var scraptText = document.DocumentNode.LastChild.FirstChild;
// var aa = scraptText.FirstChild;
Assert.AreEqual(10, scraptText.StreamPosition);
Assert.AreEqual(2, scraptText.Line);
Assert.AreEqual(8, scraptText.LinePosition);
}
{
var document = new Candy.HtmlDocument();
document.LoadHtml(@"
<script>foo</script>");
var scriptText = document.DocumentNode.LastChild.FirstChild;
Assert.AreEqual(10, scriptText.StreamPosition);
Assert.AreEqual(2, scriptText.Line);
Assert.AreEqual(8, scriptText.LinePosition);
}
}
public void AttributeValue()
{
{
Assert.AreEqual(""'", WebUtility.HtmlEncode("\"'"));
}
{
var input = "<div>z</div>";
var doc = new Candy.HtmlDocument();
doc.BackwardCompatibility = true;
doc.LoadHtml(input);
var divNode = doc.DocumentNode.ChildNodes[0];
divNode.Attributes.Add("name", "value1value2");
Assert.AreEqual("value1value2", divNode.Attributes[0].Value);
Assert.AreEqual("<div name=\"value1value2\">z</div>", doc.DocumentNode.InnerHtml);
}
{
var input = "<div>z</div>";
var doc = new Candy.HtmlDocument();
doc.BackwardCompatibility = true;
doc.LoadHtml(input);
var divNode = doc.DocumentNode.ChildNodes[0];
divNode.Attributes.Add("name", "value1\"value2");
Assert.AreEqual("value1\"value2", divNode.Attributes[0].Value);
Assert.AreEqual("<div name=\"value1"value2\">z</div>", doc.DocumentNode.InnerHtml);
}
{
var input = "<div>z</div>";
var doc = new Candy.HtmlDocument();
doc.BackwardCompatibility = true;
doc.LoadHtml(input);
var divNode = doc.DocumentNode.ChildNodes[0];
divNode.Attributes.Add("name", "value1"value2");
Assert.AreEqual("value1"value2", divNode.Attributes[0].Value);
Assert.AreEqual("<div name=\"value1"value2\">z</div>", doc.DocumentNode.InnerHtml);
}
{
var input = "<div>z</div>";
var doc = new Candy.HtmlDocument();
doc.BackwardCompatibility = true;
doc.LoadHtml(input);
var divNode = doc.DocumentNode.ChildNodes[0];
divNode.Attributes.Add("name", "value1'value2");
divNode.Attributes[0].QuoteType = AttributeValueQuote.SingleQuote;
Assert.AreEqual("value1'value2", divNode.Attributes[0].Value);
Assert.AreEqual("<div name='value1'value2'>z</div>", doc.DocumentNode.InnerHtml);
}
{
var input = "<div>z</div>";
var doc = new Candy.HtmlDocument();
doc.BackwardCompatibility = true;
doc.LoadHtml(input);
var divNode = doc.DocumentNode.ChildNodes[0];
divNode.Attributes.Add("name", "value1'value2");
divNode.Attributes[0].QuoteType = AttributeValueQuote.SingleQuote;
Assert.AreEqual("value1'value2", divNode.Attributes[0].Value);
Assert.AreEqual("<div name='value1'value2'>z</div>", doc.DocumentNode.InnerHtml);
}
{
var input = "<div>z</div>";
var doc = new Candy.HtmlDocument();
doc.BackwardCompatibility = false;
doc.LoadHtml(input);
var divNode = doc.DocumentNode.ChildNodes[0];
divNode.Attributes.Add("name", "value1value2");
Assert.AreEqual("value1value2", divNode.Attributes[0].Value);
Assert.AreEqual("<div name=\"value1value2\">z</div>", doc.DocumentNode.InnerHtml);
}
{
var input = "<div>z</div>";
var doc = new Candy.HtmlDocument();
doc.BackwardCompatibility = false;
doc.LoadHtml(input);
var divNode = doc.DocumentNode.ChildNodes[0];
divNode.Attributes.Add("name", "value1\"value2");
Assert.AreEqual("value1\"value2", divNode.Attributes[0].Value);
Assert.AreEqual("<div name=\"value1"value2\">z</div>", doc.DocumentNode.InnerHtml);
}
{
var input = "<div>z</div>";
var doc = new Candy.HtmlDocument();
doc.BackwardCompatibility = false;
doc.LoadHtml(input);
var divNode = doc.DocumentNode.ChildNodes[0];
divNode.Attributes.Add("name", "value1"value2");
Assert.AreEqual("value1"value2", divNode.Attributes[0].Value);
Assert.AreEqual("<div name=\"value1"value2\">z</div>", doc.DocumentNode.InnerHtml);
}
{
var input = "<div>z</div>";
var doc = new Candy.HtmlDocument();
doc.BackwardCompatibility = false;
doc.LoadHtml(input);
var divNode = doc.DocumentNode.ChildNodes[0];
divNode.Attributes.Add("name", "value1'value2");
divNode.Attributes[0].QuoteType = AttributeValueQuote.SingleQuote;
Assert.AreEqual("value1'value2", divNode.Attributes[0].Value);
Assert.AreEqual("<div name='value1'value2'>z</div>", doc.DocumentNode.InnerHtml);
}
{
var input = "<div>z</div>";
var doc = new Candy.HtmlDocument();
doc.BackwardCompatibility = false;
doc.LoadHtml(input);
var divNode = doc.DocumentNode.ChildNodes[0];
divNode.Attributes.Add("name", "value1'value2");
divNode.Attributes[0].QuoteType = AttributeValueQuote.SingleQuote;
Assert.AreEqual("value1'value2", divNode.Attributes[0].Value);
Assert.AreEqual("<div name='value1'value2'>z</div>", doc.DocumentNode.InnerHtml);
}
}