public void BasicWithImagesTest()
{
string h = "<div><p><img src='http://example.com/' alt=Image></p><p><img src='ftp://ftp.example.com'></p></div>";
string cleanHtml = Dcsoup.Clean(h, Whitelist.BasicWithImages);
Assert.AreEqual("<p><img src=\"http://example.com/\" alt=\"Image\"></p><p><img></p>", TextUtil.StripNewlines(cleanHtml));
}
public void NoImplicitFormForTextAreas()
{
// old jsoup parser would create implicit forms for form children like <textarea>, but no more
Document doc = Dcsoup.Parse("<textarea>One</textarea>");
Assert.AreEqual("<textarea>One</textarea>", doc.Body.Html);
}
public void TestHtmlAndXmlSyntax()
{
string h = "<!DOCTYPE html><body><img async checked='checked' src='&<>\"'><>&"<foo />bar";
Document doc = Dcsoup.Parse(h);
doc.OutputSettings.Syntax = DocumentSyntax.Html;
Assert.AreEqual("<!DOCTYPE html>\n" +
"<html>\n" +
" <head></head>\n" +
" <body>\n" +
" <img async checked src=\"&<>"\"><>&\"\n" +
" <foo />bar\n" +
" </body>\n" +
"</html>", doc.Html);
doc.OutputSettings.Syntax = DocumentSyntax.Xml;
Assert.AreEqual("<!DOCTYPE html>\n" +
"<html>\n" +
" <head></head>\n" +
" <body>\n" +
" <img async=\"\" checked=\"checked\" src=\"&<>"\" /><>&\"\n" +
" <foo />bar\n" +
" </body>\n" +
"</html>", doc.Html);
}
public void TestAFlowContents()
{
// html5 has <a> as either phrasing or block
Document doc = Dcsoup.Parse("<a>Hello <div>there</div> <span>now</span></a>");
Assert.AreEqual("<a>Hello <div>there</div> <span>now</span></a>", TextUtil.StripNewlines(doc.Body.Html));
}
public void CommentBeforeHtml()
{
string h = "<!-- comment --><!-- comment 2 --><p>One</p>";
Document doc = Dcsoup.Parse(h);
Assert.AreEqual("<!-- comment --><!-- comment 2 --><html><head></head><body><p>One</p></body></html>", TextUtil.StripNewlines(doc.Html));
}
public void NormalisedBodyAfterContent()
{
Document doc = Dcsoup.Parse("<font face=Arial><body class=name><div>One</div></body></font>");
Assert.AreEqual("<html><head></head><body class=\"name\"><font face=\"Arial\"><div>One</div></font></body></html>",
TextUtil.StripNewlines(doc.Html));
}
public void TestSpanContents()
{
// like h1 tags, the spec says SPAN is phrasing only, but browsers and publisher treat span as a block tag
Document doc = Dcsoup.Parse("<span>Hello <div>there</div> <span>now</span></span>");
Assert.AreEqual("<span>Hello <div>there</div> <span>now</span></span>", TextUtil.StripNewlines(doc.Body.Html));
}
public Document GetAsDoc()
{
if (htmlDoc is null)
{
string baseUri = HttpUtil.GetBaseUrl(Endpoint);
using (MemoryStream memStream = new MemoryStream()) {
byte[] buffer = Encoding.UTF8.GetBytes(GetResponseContent());
memStream.Write(buffer, 0, buffer.Length);
htmlDoc = Dcsoup.Parse(memStream, "UTF-8", baseUri, Parser.HtmlParser);
}
/*
* try {
* doc = Jsoup.parse(new String(html.getBytes(), "UTF-8"), baseUri);
* doc.outputSettings().charset("UTF-8");
*
* } catch (UnsupportedEncodingException ex2) {
* ex2.printStackTrace();
* doc = Jsoup.parse(html, baseUri);
* }
*/
}
return(htmlDoc);
}
public void HandleAbsOnLocalhostFileUris()
{
Document doc = Dcsoup.Parse("<a href='password'>One/a><a href='/var/log/messages'>Two</a>", "file://localhost/etc/");
Element one = doc.Select("a").First;
Assert.AreEqual("file://localhost/etc/password", one.AbsUrl("href"));
}
public void TestDropScript()
{
string h = "<SCRIPT SRC=//ha.ckers.org/.j><SCRIPT>alert(/XSS/.source)</SCRIPT>";
string cleanHtml = Dcsoup.Clean(h, Whitelist.Relaxed);
Assert.AreEqual("", cleanHtml);
}
public void TestDropImageScript()
{
string h = "<IMG SRC=\"javascript:alert('XSS')\">";
string cleanHtml = Dcsoup.Clean(h, Whitelist.Relaxed);
Assert.AreEqual("<img>", cleanHtml);
}
public void TestDropXmlProc()
{
string h = "<?import namespace=\"xss\"><p>Hello</p>";
string cleanHtml = Dcsoup.Clean(h, Whitelist.Relaxed);
Assert.AreEqual("<p>Hello</p>", cleanHtml);
}
public void TestDropComments()
{
string h = "<p>Hello<!-- no --></p>";
string cleanHtml = Dcsoup.Clean(h, Whitelist.Relaxed);
Assert.AreEqual("<p>Hello</p>", cleanHtml);
}
public void TestRelaxed()
{
string h = "<h1>Head</h1><table><tr><td>One<td>Two</td></tr></table>";
string cleanHtml = Dcsoup.Clean(h, Whitelist.Relaxed);
Assert.AreEqual("<h1>Head</h1><table><tbody><tr><td>One</td><td>Two</td></tr></tbody></table>", TextUtil.StripNewlines(cleanHtml));
}
public void HandlesInvalidStartTags()
{
string h = "<div>Hello < There <&></div>"; // parse to <div {#text=Hello < There <&>}>
Document doc = Dcsoup.Parse(h);
Assert.AreEqual("Hello < There <&>", doc.Select("div").First.Text);
}
public void TestWithSupplementaryCharacter()
{
Document doc = Dcsoup.Parse(char.ConvertFromUtf32(135361));
TextNode t = doc.Body.TextNodes[0];
Assert.AreEqual(char.ConvertFromUtf32(135361), t.OuterHtml.Trim());
}
public void NormalisesHeadlessBody()
{
Document doc = Dcsoup.Parse("<html><body><span class=\"foo\">bar</span>");
Assert.AreEqual("<html><head></head><body><span class=\"foo\">bar</span></body></html>",
TextUtil.StripNewlines(doc.Html));
}
public void HandlesXmlDeclarationAsBogusComment()
{
string html = "<?xml encoding='UTF-8' ?><body>One</body>";
Document doc = Dcsoup.Parse(html);
Assert.AreEqual("<!--?xml encoding='UTF-8' ?--> <html> <head></head> <body> One </body> </html>", StringUtil.NormaliseWhitespace(doc.OuterHtml));
}
public void TestHgroup()
{
// jsoup used to not allow hroup in h{n}, but that's not in spec, and browsers are OK
Document doc = Dcsoup.Parse("<h1>Hello <h2>There <hgroup><h1>Another<h2>headline</hgroup> <hgroup><h1>More</h1><p>stuff</p></hgroup>");
Assert.AreEqual("<h1>Hello </h1><h2>There <hgroup><h1>Another</h1><h2>headline</h2></hgroup> <hgroup><h1>More</h1><p>stuff</p></hgroup></h2>", TextUtil.StripNewlines(doc.Body.Html));
}
public void HandlesTagsInTextarea()
{
string html = "<textarea><p>Jsoup</p></textarea>";
Document doc = Dcsoup.Parse(html);
Assert.AreEqual("<textarea><p>Jsoup</p></textarea>", doc.Body.Html);
}
public void TestNoImagesInNoScriptInHead()
{
// jsoup used to allow, but against spec if parsing with noscript
Document doc = Dcsoup.Parse("<html><head><noscript><img src='foo'></noscript></head><body><p>Hello</p></body></html>");
Assert.AreEqual("<html><head><noscript></noscript></head><body><img src=\"foo\"><p>Hello</p></body></html>", TextUtil.StripNewlines(doc.Html));
}
public void HandlesTextAfterData()
{
string h = "<html><body>pre <script>inner</script> aft</body></html>";
Document doc = Dcsoup.Parse(h);
Assert.AreEqual("<html><head></head><body>pre <script>inner</script> aft</body></html>", TextUtil.StripNewlines(doc.Html));
}
public void TestFontFlowContents()
{
// html5 has no definition of <font>; often used as flow
Document doc = Dcsoup.Parse("<font>Hello <div>there</div> <span>now</span></font>");
Assert.AreEqual("<font>Hello <div>there</div> <span>now</span></font>", TextUtil.StripNewlines(doc.Body.Html));
}
public void HandlesWhatWgExpensesTableExample()
{
// http://www.whatwg.org/specs/web-apps/current-work/multipage/tabular-data.html#examples-0
Document doc = Dcsoup.Parse("<table> <colgroup> <col> </colgroup><colgroup> <col> <col> <col> </colgroup><thead> <tr> <th> </th><th>2008 </th><th>2007 </th><th>2006 </th></tr></thead><tbody> <tr> <th scope=\"rowgroup\"> Research and development </th><td> $ 1,109 </td><td> $ 782 </td><td> $ 712 </td></tr><tr> <th scope=\"row\"> Percentage of net sales </th><td> 3.4% </td><td> 3.3% </td><td> 3.7% </td></tr></tbody><tbody> <tr> <th scope=\"rowgroup\"> Selling, general, and administrative </th><td> $ 3,761 </td><td> $ 2,963 </td><td> $ 2,433 </td></tr><tr> <th scope=\"row\"> Percentage of net sales </th><td> 11.6% </td><td> 12.3% </td><td> 12.6% </td></tr></tbody></table>");
Assert.AreEqual("<table> <colgroup> <col> </colgroup><colgroup> <col> <col> <col> </colgroup><thead> <tr> <th> </th><th>2008 </th><th>2007 </th><th>2006 </th></tr></thead><tbody> <tr> <th scope=\"rowgroup\"> Research and development </th><td> $ 1,109 </td><td> $ 782 </td><td> $ 712 </td></tr><tr> <th scope=\"row\"> Percentage of net sales </th><td> 3.4% </td><td> 3.3% </td><td> 3.7% </td></tr></tbody><tbody> <tr> <th scope=\"rowgroup\"> Selling, general, and administrative </th><td> $ 3,761 </td><td> $ 2,963 </td><td> $ 2,433 </td></tr><tr> <th scope=\"row\"> Percentage of net sales </th><td> 11.6% </td><td> 12.3% </td><td> 12.6% </td></tr></tbody></table>", TextUtil.StripNewlines(doc.Body.Html));
}
public void EmptyTdTag()
{
string h = "<table><tr><td>One</td><td id='2' /></tr></table>";
Document doc = Dcsoup.Parse(h);
Assert.AreEqual("<td>One</td>\n<td id=\"2\"></td>", doc.Select("tr").First.Html);
}
public void NoTableDirectInTable()
{
Document doc = Dcsoup.Parse("<table> <td>One <td><table><td>Two</table> <table><td>Three");
Assert.AreEqual("<table> <tbody><tr><td>One </td><td><table><tbody><tr><td>Two</td></tr></tbody></table> <table><tbody><tr><td>Three</td></tr></tbody></table></td></tr></tbody></table>",
TextUtil.StripNewlines(doc.Body.Html));
}
public void HandlesCommentsInTable()
{
string html = "<table><tr><td>text</td><!-- Comment --></tr></table>";
Document node = Dcsoup.ParseBodyFragment(html);
Assert.AreEqual("<html><head></head><body><table><tbody><tr><td>text</td><!-- Comment --></tr></tbody></table></body></html>", TextUtil.StripNewlines(node.OuterHtml));
}
public void IgnoresDupeEndTrTag()
{
Document doc = Dcsoup.Parse("<table><tr><td>One</td><td><table><tr><td>Two</td></tr></tr></table></td><td>Three</td></tr></table>"); // two </tr></tr>, must ignore or will close table
Assert.AreEqual("<table><tbody><tr><td>One</td><td><table><tbody><tr><td>Two</td></tr></tbody></table></td><td>Three</td></tr></tbody></table>",
TextUtil.StripNewlines(doc.Body.Html));
}
public void TestXhtmlReferences()
{
Document doc = Dcsoup.Parse("< > & " ' ×");
doc.OutputSettings.EscapeMode = DocumentEscapeMode.Xhtml;
Assert.AreEqual("< > & \" ' ×", doc.Body.Html);
}
public void TestScriptTagInWhiteList()
{
Whitelist whitelist = Whitelist.Relaxed;
whitelist.AddTags("script");
Assert.IsTrue(Dcsoup.IsValid("Hello<script>alert('Doh')</script>World !", whitelist));
}
