public void SupplyOutputSettings()
{
// test that one can override the default document output settings
DocumentOutputSettings os = new DocumentOutputSettings();
os.PrettyPrint = false;
os.EscapeMode = DocumentEscapeMode.Extended;
os.Charset = Encoding.ASCII;
string html = "<div><p>ℬ</p></div>";
string customOut = Dcsoup.Clean(html, "http://foo.com/", Whitelist.Relaxed, os);
string defaultOut = Dcsoup.Clean(html, "http://foo.com/", Whitelist.Relaxed);
Assert.AreNotEqual(defaultOut, customOut);
Assert.AreEqual("<div><p>ℬ</p></div>", customOut);
Assert.AreEqual("<div>\n" +
" <p>ℬ</p>\n" +
"</div>", defaultOut);
os.Charset = Encoding.ASCII;
os.EscapeMode = DocumentEscapeMode.Base;
string customOut2 = Dcsoup.Clean(html, "http://foo.com/", Whitelist.Relaxed, os);
Assert.AreEqual("<div><p>ℬ</p></div>", customOut2);
}
/// <summary>
/// Set the current output syntax of the specified document settings, and returns the document settings itself.
/// </summary>
/// <param name="self">
/// The input <see cref="DocumentOutputSettings"/>,
/// which acts as the <b>this</b> instance for the extension method.
/// </param>
/// <param name="syntax">the new syntax to use</param>
/// <returns>The input <see cref="DocumentOutputSettings"/>, for method chaining.</returns>
/// <seealso cref="DocumentOutputSettings.Syntax">DocumentOutputSettings.Syntax</seealso>
public static DocumentOutputSettings Syntax(this DocumentOutputSettings self, DocumentSyntax syntax)
{
self.Syntax = syntax;
return self;
}
/// <summary>
/// Get safe HTML from untrusted input HTML,
/// by parsing input HTML and filtering it through a white-list of permitted tags and attributes.
/// </summary>
/// <param name="bodyHtml">input untrusted HTML (body fragment)</param>
/// <param name="baseUri">URL to resolve relative URLs against</param>
/// <param name="whitelist">white-list of permitted HTML elements</param>
/// <param name="outputSettings">document output settings; use to control pretty-printing and entity escape modes</param>
/// <returns>safe HTML (body fragment)</returns>
/// <seealso cref="Supremes.Safety.Cleaner.Clean(Supremes.Nodes.Document)">Supremes.Safety.Cleaner.Clean(Supremes.Nodes.Document)</seealso>
public static string Clean(string bodyHtml, string baseUri, Whitelist whitelist, DocumentOutputSettings outputSettings)
{
Document dirty = ParseBodyFragment(bodyHtml, baseUri);
Cleaner cleaner = new Cleaner(whitelist);
Document clean = cleaner.Clean(dirty);
clean.OutputSettings = outputSettings;
return(clean.Body.Html);
}
/// <summary>
/// Set the current output charset of the specified document settings, and returns the document settings itself.
/// </summary>
/// <param name="self">
/// The input <see cref="DocumentOutputSettings"/>,
/// which acts as the <b>this</b> instance for the extension method.
/// </param>
/// <param name="charset">the new charset to use</param>
/// <returns>The input <see cref="DocumentOutputSettings"/>, for method chaining.</returns>
/// <seealso cref="DocumentOutputSettings.Charset">DocumentOutputSettings.Charset</seealso>
public static DocumentOutputSettings Charset(this DocumentOutputSettings self, Encoding charset)
{
self.Charset = charset;
return self;
}
/// <summary>
/// Set the current output charset of the specified document settings, and returns the document settings itself.
/// </summary>
/// <param name="self">
/// The input <see cref="DocumentOutputSettings"/>,
/// which acts as the <b>this</b> instance for the extension method.
/// </param>
/// <param name="charset">the new charset name to use</param>
/// <returns>The input <see cref="DocumentOutputSettings"/>, for method chaining.</returns>
/// <seealso cref="DocumentOutputSettings.Charset">DocumentOutputSettings.Charset</seealso>
public static DocumentOutputSettings Charset(this DocumentOutputSettings self, String charset)
{
self.Charset = Encoding.GetEncoding(charset);
return self;
}
/// <summary>
/// Set the current HTML escape mode of the specified document settings, and returns the document settings itself.
/// </summary>
/// <param name="self">
/// The input <see cref="DocumentOutputSettings"/>,
/// which acts as the <b>this</b> instance for the extension method.
/// </param>
/// <param name="escapeMode">the new escape mode to use</param>
/// <returns>The input <see cref="DocumentOutputSettings"/>, for method chaining.</returns>
/// <seealso cref="DocumentOutputSettings.EscapeMode">DocumentOutputSettings.EscapeMode</seealso>
public static DocumentOutputSettings EscapeMode(this DocumentOutputSettings self, DocumentEscapeMode escapeMode)
{
self.EscapeMode = escapeMode;
return self;
}
/// <summary>
/// Set the specified document's current output settings, and returns the document itself.
/// </summary>
/// <param name="self">
/// The input <see cref="Document"/>,
/// which acts as the <b>this</b> instance for the extension method.
/// </param>
/// <param name="outputSettings">new output settings</param>
/// <returns>The input <see cref="Document"/>, for method chaining.</returns>
/// <seealso cref="Document.OutputSettings">Document.OutputSettings</seealso>
public static Document OutputSettings(this Document self, DocumentOutputSettings outputSettings)
{
self.OutputSettings = outputSettings;
return self;
}
/// <summary>
/// Set the current tag indent amount of the specified document settings, used when pretty printing, and returns the document settings itself.
/// </summary>
/// <param name="self">
/// The input <see cref="DocumentOutputSettings"/>,
/// which acts as the <b>this</b> instance for the extension method.
/// </param>
/// <param name="indentAmount">number of spaces to use for indenting each level</param>
/// <returns>The input <see cref="DocumentOutputSettings"/>, for method chaining.</returns>
/// <seealso cref="DocumentOutputSettings.IndentAmount">DocumentOutputSettings.IndentAmount</seealso>
public static DocumentOutputSettings IndentAmount(this DocumentOutputSettings self, int indentAmount)
{
self.IndentAmount = indentAmount;
return self;
}
/// <summary>
/// Set if outline mode is enabled to the specified document settings, and returns the document settings itself.
/// </summary>
/// <param name="self">
/// The input <see cref="DocumentOutputSettings"/>,
/// which acts as the <b>this</b> instance for the extension method.
/// </param>
/// <param name="outlineMode">the new outline mode</param>
/// <returns>The input <see cref="DocumentOutputSettings"/>, for method chaining.</returns>
/// <seealso cref="DocumentOutputSettings.Outline">DocumentOutputSettings.Outline</seealso>
public static DocumentOutputSettings Outline(this DocumentOutputSettings self, bool outlineMode)
{
self.Outline = outlineMode;
return self;
}
/// <summary>
/// Set if pretty printing is enabled to the specified document settings, and returns the document settings itself.
/// </summary>
/// <param name="self">
/// The input <see cref="DocumentOutputSettings"/>,
/// which acts as the <b>this</b> instance for the extension method.
/// </param>
/// <param name="pretty">the new pretty print setting</param>
/// <returns>The input <see cref="DocumentOutputSettings"/>, for method chaining.</returns>
/// <seealso cref="DocumentOutputSettings.PrettyPrint">DocumentOutputSettings.PrettyPrint</seealso>
public static DocumentOutputSettings PrettyPrint(this DocumentOutputSettings self, bool pretty)
{
self.PrettyPrint = pretty;
return self;
}
